Zentrum für Entwicklungsforschung
Water, Sanitation and Agriculture
Linkages and Impacts on Health and Nutrition Outcomes in Rural Ethiopia
Inaugural – Dissertation
Zur
Erlangung des Grades
Doktor der Agrarwissenschaften (Dr.agr.)
der
Landwirtschaftlichen Fakultät
der
Rheinischen Friedrich-Wilhelms-Universität Bonn
von
Muhammed Abdella Usman
aus
Nefas Mewcha, Äthiopien
Bonn 2017
i
Zentrum für Entwicklungsforschung
Referent: Professor Dr. Joachim von Braun
Korreferent: Professor Dr. Jan Börner
Tag der mündlichen Prüfung: 20 .07. 2017
ii
Abstract
In rural areas access to improved water supply is inadequate; consequently, most households rely on unimproved water sources, including unprotected wells/springs, streams, and surface water, which are easily polluted by human and animal feces. Moreover, irrigated agriculture has complex interactions with water supply and sanitation (WATSAN) services as separate sources of water for drinking and for agricultural use do not exist in several areas. For this analysis, a household survey has been conducted in rural areas of Fogera and Mecha districts of Ethiopia between February and June 2014. A sample of 454 agricultural households was randomly selected using a stratified two-stage cluster sampling method.
The survey collected a range of information including anthropometric measures for under-five children to examine child nutritional status. In addition, assessment of the microbial quality of stored drinking water and community water sources were undertaken. The number of Escherichia coli (E.coli) colony-forming units per 100ml water was used as an indicator of fecal contamination, and the results demonstrate that 58 percent of household stored water samples and 74 percent of water sources were contaminated with E.coli. Our results also show that uncontaminated household storage water and safe child stool disposal decrease incidence of child diarrhea by 15 and 23 percentage points, respectively. In contrast, neighborhood concentration of pit latrine increases the incidence of child diarrhea by 13 percentage points. The latter result casts serious doubt on the assumed health and social benefits of moving from open to fixed-location defecation. Creating open defecation free communities in rural areas is not enough to achieve the desired health benefits of sanitation. To protect rural households from the risk of contracting communicable diseases, existing pit latrines should be upgraded to make them safer to use –fly-proofed and hygienic.
Using anthropometric measures of under-five children, our results show that WATSAN services
are strongly associated with improved weight-for-age z-score but its correlation with height-for-
age z-score is not statistically significant at any of the conventional levels. Dietary diversity of
child feeding practice and health indicator variables, such as number of antenatal care visits and
delivery with a health professional, turn out to be strong predictors of both nutritional outcome
measures. On the other hand, although the domestic use of irrigation water significantly
increases household’s overall morbidity status except for diarrhea, it substantially reduces the
burden of time spend on water collection for women. Recognizing the multiple-use of water
resources in rural areas where access to improved water supply is inadequate or where there is
only one water source for all household needs would be vital to design the right intervention.
Promoting improved WATSAN from ‘farm to fork’ through water source protection, appropriate
water-lifting technologies, ensuring households do not use irrigation runoff for drinking, and
adopting household water treatment and safe storage to reduce microbiological contamination
so that the health risks associated with domestic use of irrigation water may be minimized.
iii
Zusammenfassung
In ländlichen Gebieten ist der Zugang zu einer verbesserten Wasserversorgung unzureichend. Infolgedessen sind die meisten Haushalte auf Wasser aus hygienisch bedenklichen Quellen, einschließlich ungeschützten Brunnen/Quellen, Strömen und Oberflächenwasser, die leicht durch menschliche und tierische Fäkalien verschmutzt werden, angewiesen. Darüber hinaus bestehen komplexe Wechselwirkungen zwischen der bewässerten Landwirtschaft und der Wasserversorgung und Abwasserentsorgung (WATSAN), da getrennte Wasserquellen für Trink- und Landwirtschaftszwecke in mehreren Regionen nicht existieren. Für diese Analyse wurde in den ländlichen Gebieten der Bezirke Fogera und Mecha in Äthiopien zwischen Februar und Juni 2014 eine Haushaltsbefragung durchgeführt. Eine Stichprobe von 454 landwirtschaftlichen Haushalten wurde nach dem Zufallsprinzip mittels einer Klumpenstichprobe (zweistufig, geschichtet) ausgewählt.
Mittels einer Umfrage wurde eine Reihe von Informationen gesammelt, einschließlich anthropometrischer Messungen für Kinder unter fünf Jahren um deren Ernährungsstatus zu prüfen. Zusätzlich wurde eine Bewertung der mikrobiellen Qualität des gespeicherten Trinkwassers und der gemeinschaftlichen Wasserquellen vorgenommen. Die Anzahl der Escherichia coli (E.coli) koloniebildenden Einheiten pro 100ml Wasser wurde als ein Indikator für die fäkale Verunreinigung verwendet. Die Ergebnisse zeigen auf, dass 58 Prozent der Haushaltswasserproben und 74 Prozent der Wasserquellen mit E.coli kontaminiert waren. Unsere Ergebnisse zeigen zudem, dass unkontaminiertes Haushaltswasser und eine sichere Fäkalienentsorgung bei Kindern das Auftreten von Kinderdurchfall um jeweils 15 beziehungsweise 23 Prozentpunkte verringern. Im Gegensatz dazu erhöht die Konzentration von Grubenlatrinen in der Nachbarschaft die Häufigkeit von Durchfall bei Kindern um 13 Prozentpunkte. Das letztgenannte Ergebnis führt zu ernsthaften Zweifeln an den bislang angenommenen gesundheitlichen und sozialen Vorteilen des Umstiegs von öffentlicher zu stationärer Defäkation. Die Errichtung von Gemeinschaften in ländlichen Gebieten in denen Stuhlgang nicht mehr in der Öffentlichkeit praktiziert wird, reicht nicht aus um die gewünschten gesundheitlichen Vorteile von Sanitärmaßnahmen zu erreichen. Um ländliche Haushalte vor dem Risiko der Ansteckung übertragbarer Krankheiten zu schützen, sollten bestehende Grubenlatrinen ausgebaut werden, um sie sicherer zu machen – geschützt vor Fliegen und hygienisch.
Die Ergebnisse der anthropometrischen Messungen von Kindern unter fünf Jahren zeigen, dass WATSAN-Dienstleistungen stark mit einem verbesserten „weight-for-age“-Standardisierung verbunden sind, jedoch ist ihre Korrelation mit der „height-for-age“-Standardisierung nicht statistisch signifikant. Eine ausgewogene Kinderernährung und Gesundheitsindikatoren, wie zum Beispiel die Anzahl der vorgeburtlichen Betreuungsbesuche und die von medizinisch betreuten Geburten, erweisen sich als starke Einflusswerte für beide Indikatoren des Ernährungszustands. Obwohl die Morbidität, außer für Durchfall, durch die häusliche Benutzung von Bewässerungswasser insgesamt signifikant erhöht wird, wird auf der anderen Seite die Last der Frauen hinsichtlich Zeit, die für die Wassersammlung gebraucht wird, erheblich reduziert. In Anbetracht der Mehrfachnutzung der Wasserressourcen in ländlichen Gebieten, in denen der Zugang zu einer verbesserten Wasserversorgung nicht ausreichend ist oder nur eine Wasserquelle für alle Haushaltsbedürfnisse vorhanden ist, wäre es entscheidend, die richtige Maßnahme durchzuführen. Das beinhaltet die Förderung von verbesserter Wasser – und Sanitärversorgung auf dem Feld und den Haushalten durch den Schutz von Wasserquellen, geeignete Wasserförderungstechnologien, das Sicherstellen, dass private Haushalte nicht den Abfluss von Bewässerungswasser zum Trinken verwenden und die Übernahme von „Household water treatment and safe storage“-Methoden zur Verringerung der mikrobiologischen Kontamination. Mit dem Ziel, die Gesundheitsrisiken im Zusammenhang mit der häuslichen Nutzung von Bewässerungswasser, zu minimieren.
iv
Table of Contents
Abstract ................................................................................................................................ ii
Zusammenfassung ................................................................................................................ iii
List of Figures ....................................................................................................................... vi
List of Tables ....................................................................................................................... vii
List of Acronyms ................................................................................................................... ix
Acknowledgements ............................................................................................................... x
1. General Introduction ...................................................................................................... 1
1.1 Background Information .................................................................................................. 2
1.2 Motivation and Problem Statement ................................................................................ 4
1.3 Objectives ......................................................................................................................... 7
1.4 Significance of the Study .................................................................................................. 8
1.5 Organization of the Thesis ............................................................................................... 9
2. Data Collection and Survey Descriptions ....................................................................... 10
2.1 Description of the Study Areas ...................................................................................... 10
2.2 Data Collection and Organization of the Survey ............................................................ 13
2.3 Survey Design, Sampling Frame and Sample Size .......................................................... 14
2.3.1 Questionnaires and ethical consideration .............................................................. 15
2.3.2 Anthropometric measures and water sample testing ............................................ 16
2.3.3 Follow-up survey ..................................................................................................... 17
3. Drivers of Microbial Quality of Household Drinking Water in Rural Households ............. 18
3.1 Introduction.................................................................................................................... 18
3.2 Context and Related Literature ...................................................................................... 20
3.3 Methods and Data .......................................................................................................... 24
3.3.1 Data ......................................................................................................................... 24
3.3.2 Empirical estimations .............................................................................................. 28
3.4 Results and Discussion ................................................................................................... 28
3.4.1 Bivariate analysis..................................................................................................... 28
3.4.2 Multivariate analysis ............................................................................................... 32
3.5 Conclusions and Recommendations .............................................................................. 39
4. The Impact of Drinking Water Quality and Sanitation on Child Health ........................... 42
4.1 Introduction.................................................................................................................... 42
4.2 Water Supply, Sanitation, Hygiene and Health .............................................................. 44
4.2.1 Water supply and health......................................................................................... 45
4.2.2 Sanitation and health .............................................................................................. 47
4.2.3 Hygiene behavior and health .................................................................................. 48
4.3 Data and Methods .......................................................................................................... 50
4.3.1 Variables and measurement ................................................................................... 50
4.3.2 Descriptive statistics ............................................................................................... 52
4.4 Empirical Strategy .......................................................................................................... 58
v
4.4.1 Instrumental variable approach ............................................................................. 59
4.4.2 Bivariate probit estimator ....................................................................................... 59
4.5 Estimation Results and Discussion ................................................................................. 61
4.6 Conclusions and Recommendations .............................................................................. 72
4.7 A Cost-Benefit Analysis (CBA) for Rural Water Supply in Ethiopia ................................ 75
Appendix 4 .......................................................................................................................... 82
5. Water, sanitation, and agriculture: Linkages and impacts on health outcomes .............. 83
5.1 Introduction.................................................................................................................... 83
5.2 Linkage between WASH, Health, and Agriculture.......................................................... 84
5.3 Theoretical Framework .................................................................................................. 85
5.4 Data and Empirical Strategy ........................................................................................... 88
5.4.1 Data ......................................................................................................................... 88
5.4.2 Empirical strategy ................................................................................................... 90
5.5 Results and Discussion ................................................................................................... 91
5.5.1 The influence of demographic and WATSAN characteristics on diarrhea and other water-related diseases .......................................................................................................... 91
5.5.1.1 The interaction between irrigation and domestic water supply ..................... 95
5.5.1.2 Estimation results: Impact of irrigation on water quantity and diarrhea ....... 99
5.5.1.3 Estimation results: Domestic use of irrigation water and household water quality, quantity, and health outcomes ........................................................................... 102
5.6 Conclusions and Recommendations ............................................................................ 106
6. Determinants of Child Nutrition Outcomes ................................................................. 108
6.1 Introduction.................................................................................................................. 108
6.2 WASH and Child Nutrition Outcomes .......................................................................... 110
6.3 Data and Methods ........................................................................................................ 112
6.4 Results and Discussion ................................................................................................. 113
6.4.1 Multivariate regression results ............................................................................. 118
6.5 Conclusion and Implications ........................................................................................ 124
7. General Conclusions and Policy Recommendations ..................................................... 125
7.1 Potential Intervention Areas ........................................................................................ 126
7.1.1 Rural water supply ................................................................................................ 126
7.1.2 Sanitation infrastructure and practices ................................................................ 128
7.1.3 Hygiene behavior awareness ................................................................................ 129
7.1.4 Small-scale irrigation agriculture .......................................................................... 130
7.2 Limitations of the Study and Future Research Directions ........................................... 131
8. References ................................................................................................................. 133
Appendix A ........................................................................................................................ 147
Appendix B ........................................................................................................................ 157
vi
List of Figures
Figure 1-1: The linkage between agriculture and nutrition and health ......................................... 7
Figure 2-1: Map of the study areas ............................................................................................... 11
Figure 3-1: Percentage of population by type of drinking water source in Ethiopia, 2015 ......... 19
Figure 3-2: Major problems related to drinking water supply in the study areas ....................... 31
Figure 3-3: Household members who usually collect when water source is not on premises .... 32
Figure 4-1: Percentage of population by type of sanitation facility in Ethiopia, 2015 ................. 43
Figure 4-2: The “F-diagram” shows the pathways of fecal disease transmission and the barriers
that can prevent infection ............................................................................................................ 45
Figure 4-3: Method of disposal for the last child stool preceding the survey .............................. 54
Figure 4-4: Causes of childhood diarrhea reported by the primary caretakers ........................... 58
Figure 4-5: Prevalence of diarrhea for under-five children in the follow-up survey .................... 70
Figure 4-6: Proportion of children with diarrhea by number of episodes ................................... 71
Figure 4-7: Proportion of children with diarrhea by episode duration ........................................ 72
Figure 5-1: Prevalence of illness in the last two months by income quintile ............................... 94
Figure 5-2: Primary water sources for irrigation purpose ............................................................ 96
Figure 5-3: Primary source of drinking water on irrigation site ................................................... 97
Figure 5-4: Major problems related to irrigation water sources .................................................. 98
Figure 5-5: Irrigation water use for domestic purposes ............................................................... 99
Figure 6-1: Percentage of undernourished children and average dietary supply adequacy trend
over time ..................................................................................................................................... 108
Figure 6-2: Trends in nutritional status of under-five children, 2000-2014 ............................... 111
Figure 6-3: Correlation between different anthropometric indicators ...................................... 114
Figure 6-4: Distribution of z-scores for under-five children ....................................................... 115
vii
List of Tables
Table 1-1: Estimated drinking water and sanitation coverage 2015 update in percentage .......... 2
Table 3-1: JMP Classification of drinking water source types and sanitation facilities ................ 21
Table 3-2: Compliance of drinking water sources in Ethiopia for thermotolerant coliforms a .... 22
Table 3-3: Descriptive statistics – household and community characteristics (N=454) ............... 27
Table 3-4: Relationships between water sources, collection and stored drinking water quality 30
Table 3-5: Community water source sample test results ............................................................. 31
Table 3-6: Estimates from OLS regression predicting the natural log of E.coli ............................ 35
Table 3-7: Estimates from Logistic regression predicting E.coli contamination (1 if E.coli >= 1) . 36
Table 4-1: Transmission Routes of Water-Related Diseases ........................................................ 44
Table 4-2: A meta-analysis of percentage change in diarrheal diseases by intervention type .... 49
Table 4-3: List of variables: Definitions and summary statistics................................................... 53
Table 4-4: Diarrhea incidence by demographic characteristics (last two weeks) ........................ 55
Table 4-5: Diarrhea incidence by WASH characteristics (last two weeks) ................................... 57
Table 4-6: Household water quality: First-stage regression ......................................................... 62
Table 4-7: Health effects of water quality and sanitation: Diarrhea in under-five children ........ 63
Table 4-8: Health effects of water quality and sanitation: Diarrhea in under-three children ..... 67
Table 4-9: Health effects of water quality and sanitation: Diarrhea in under-five children ........ 68
Table 4-10: Metrics summary of value per DALY calculation ....................................................... 78
Table 4-11: Metrics and summary of value per DALY calculation ................................................ 79
Table 4-12: Summary of cost-benefit analysis .............................................................................. 80
Table 4-13: Present discounted value (PDV) computation details ............................................... 81
Table 5-1: Descriptive statistics .................................................................................................... 89
Table 5-2: Diarrhea incidence by demographic and WASH characteristics (last two months) .... 92
Table 5-3: The influence of demographic and WASH characteristics on water-related diseases and
overall morbidity (last two months) ............................................................................................. 93
Table 5-4: Types of irrigation mostly practiced by farmers .......................................................... 95
Table 5-5: Main types of water delivery system from irrigation water source ............................ 96
Table 5-6: Incidence of diseases by drinking water source on irrigation site .............................. 98
Table 5-7: Health effects of irrigation—Diarrhea ....................................................................... 101
Table 5-8: Effects of domestic use of irrigation water on diarrhea ............................................ 103
Table 6-1: WHO classification of severity of malnutrition prevalence ....................................... 111
Table 6-2: Descriptive statistics .................................................................................................. 113
Table 6-3: The prevalence of malnutrition based on the WHO 2006 reference group ............. 115
Table 6-4: Malnutrition prevalence in under-five children by demographic characteristics ..... 116
Table 6-5: Malnutrition prevalence in under-five children by WASH characteristics ................ 117
Table 6-6: Malnutrition prevalence in under-five children by income distribution ................... 118
viii
Table 6-7: Multivariate regression model for weight-for-age/underweight .............................. 120
Table 6-8: Multivariate regression model for height-for-age/stunting ...................................... 123
ix
List of Acronyms
ANRS Amhara National Regional State
AG-WATSAN Agriculture-Water and Sanitation
ATE/ATT Average Treatment Effect / Average Treatment on the Treated
CSA Central Statistical Agency
DHS Demographic and Health Survey
EEA Ethiopian Economics Association
GPS Global Positioning System
HEW Health Extension Worker
HWTS Household Water Treatment and Safe Storage
L/p/c/d Liter Per Capita Per Day
MLSB Membrane Lauryl Sulphate Broth
ORDA Organization for Rehabilitation and Development in Amhara
WHO World Health Organization
UNICEF United Nations Children’s Fund
FMoH Federal Ministry of Health
POS Point-of-Source
POU Point-of-Use
PSU Primary Sampling Unit
SDGs Sustainable Development Goals
SSU Secondary Sampling Unit
WASH Water, Sanitation and Hygiene
WASTAN Water and Sanitation
WUA Water User Association
OLS Ordinary Least Squares
CLTS Community-Led Total Sanitation
MoWE Ministry of Water and Energy
SSI Small-scale Irrigation
ZEF Center for Development Research, University of Bonn
x
Acknowledgements
In the name of Allah the most Merciful and most Compassionate
All praises are due to God alone, the Creator and Sustainer of all the worlds. He is the Almighty,
the greatest of all. Indeed, my success is only by Allah. Without the support from many important
people; however, it would have been difficult to get my thesis done. My special thanks go to my
supervisor Joachim von Braun for offering me the Ph.D. position without which I could never have
achieved this. He has given me a number of insightful feedbacks and recommendations that have
momentously helped me to improve my thesis. Moreover, his support, intellectual guidance, and
encouragement was instrumental and has shaped this work to what it is. I am also highly indebted
to Nicolas Gerber for the numerous discussions and comments throughout the writing of my
thesis. Although he is busy like anyone else, he often welcomes me whenever I show up myself
at his office. I also gratefully acknowledge Jan Börner for accepting to be my second supervisor.
I am grateful for the Bill and Melinda Gates Foundation for funding my Ph.D. study through the
“WATSAN-AGRICULTURE—guiding pro-poor investments in the nexus among domestic water
quality and quantity, sanitation and hygiene, and agriculture from the bottom-up.” I would like
to acknowledge the support from the Center for Development Research (ZEF), the German Agro
Action (Welthungerhilfe) and the Organization for Rehabilitation and Development in Amhara
(ORDA) during my field research in Ethiopia. I would also like to thank all the Ethiopian Economics
Association (EEA) staff members, especially Assefa Admassie and Samuel G/Sellasie for their
support and personal consultation to make the field work more productive and making my stay
more interesting. I would like to thank all the people who provided valuable comments and
suggestions at various seminars, workshops, and international conferences to improve my
research. I would also like to thank Samantha Antonini for her review and suggestions on the
earlier versions of my thesis. I also thank Daniel Tsegai and Evita H. Pangaribowo for their
guidance and support at the early stage of my research. I would also like to acknowledge the
Hermann Eiselen Doctoral Programme of the Foundation of fiat panis for their financial support
to present my research findings in international scientific conferences.
My appreciation goes to Günther Manske, Guido Lüchters, Kyere Vincent Narteye, Volker Merx,
Mrs. Maike Retat-Amin, and to all friends and colleagues at ZEF, especially 2012 batch. I would
also like to appreciate my friends Mekbib G. Haile, Justice Tambo and Oliver Kirui not only for
their helpful academic suggestions and discussions we have had but also the time we have spent
together playing ping pong almost every single day. I am sure that my stay at ZEF would not be
as interesting as it has been without you guys. It was too much fun and joy for the last two years.
Thanks for giving me such a joyful time. I also acknowledge all the Ph.D. students and research
assistants in the ‘AG-WATSAN’ Nexus project for their valuable comments and suggestions at
xi
several of our project seminars and workshops. I am also grateful to all the data collectors for
their hard work during the household survey.
I owe everything to my parents and siblings who encouraged and assisted me at every step of my
personal and academic life and longed to see this accomplishment come true. I would also like
to express my appreciation to my friends back home for supporting and encouraging me
throughout my study. You all have been a great moral support and inspiration.
Last but not least, my greatest gratitude goes to my family: my wife, Fatuma, and my daughters,
Tasneem and Ahlam. I was lucky to have you as my companion in my life. You are not only my
wife but also my best friend. You are my treasure and you mean a lot to me. Finding the right
words to convey how much I owe you is difficult for me. You are just amazing. Thanks for your
support all the time. I love you all dearly!
Verily, after every hardship there is relief. Alhamdulillah!!
I dedicate this work to my loving parents.
1
1. General Introduction
Worldwide more than 3.4 million people die of diseases related to water, sanitation and hygiene
(WASH) (Prüss-Üstün et al. 2008). More than 90 percent of diarrheal deaths under the age of five
years are in developing countries (UN Water 2008a). As a result of lack of access to improved
drinking water supply and sanitation (WATSAN),1 inadequate hygiene practices and malnutrition,
diarrhea is more prevalent in many developing countries. It is the second leading causes of death
among under-five children globally, which kills more children than malaria, AIDS, and measles
combined (WHO/UNICEF 2009). In most cases, these deaths could have been prevented by
providing safe drinking water, basic sanitation facilities, and promoting proper hygiene practices.
As reported by the WHO/UNICEF Joint Monitoring Programme (JMP), which monitors progress
on the water and sanitation targets, 2.5 billion people—over one-third of the world population—
lack access to improved sanitation facilities, of which a billion people practice open defecation,
and over 780 million people—80 percent of these people live in rural areas—worldwide lack
access to improved drinking water sources (WHO/UNICEF 2014).2 Most of the unserved
population lives in sub-Saharan Africa and southeastern Asian regions.
The problems of inadequate WASH services are complex and multifaceted. Globally, due to
WASH-related diseases alone, 5.5 billion productive days and 443 million school days are lost
every year (Fewtrell et al. 2007; UNDP 2006). Studies also show that provision of sanitation
facilities at school increase girl’s school enrollment rates by over 15 percent (IRC 2007; UN Water
2008b). It is estimated that improving WASH services could prevent nearly 10 percent of the
burden of disease and about 6 percent of all deaths globally (Prüss-Üstün et al. 2008).
Improved WATSAN services together with proper hygiene practices can have significant health
impacts on the population by reducing the incidence of a variety of water-related diseases. A
wide range of empirical studies show that improvement in WATSAN service affect health
outcomes, as measured by less diarrhea, improved child growth, reduction in parasitic and worm
infections, skin diseases, trachoma, and lower rates of morbidity and mortality (Checkley et al.
2004; Esrey 1996; Esrey et al. 1991; Fenn et al. 2012; Günther & Fink 2010; Jalan & Ravallion
2003). In addition to improving human health by putting a barrier for the transmission of
waterborne and other infectious diseases, improved WATSAN services can transform the lives of
1 WHO/UNICEF JMP for WATSAN defines an improved drinking water supply refer to sources that are adequately
protected from external contamination, especially from the contamination with fecal matter, but are not necessarily free from pathogens or safe for drinking. Similarly, an improved sanitation facility is one that hygienically separates human feces from human contacts.
2 Open defecation is defined as defecation in fields, behind bushes, forests, in roadside ditches, bodies of water or other open spaces.
2
many disadvantaged people in developing countries through various channels. Because
investments in WATSAN infrastructures generate a wide-ranging economic benefit.
1.1 Background Information
The coverage of improved WATSAN infrastructure in Ethiopia is one of the lowest in the world.
Table 1-1 presents the coverage of access to improved WATSAN services in Ethiopia between
1990 and 2015. About 57 percent of the population has access to improved drinking water
sources, while these numbers drop sharply in rural areas; and 58 percent of the population lacks
access to basic sanitation facilities, and nearly 30 million people—29 percent of the total
population—practicing open defecation (WHO/UNICEF 2015). These figures conceal various
inequalities between rural and urban areas in terms of access and service levels. The impacts of
inadequate WATSAN services on the country’s GDP is expected to be substantial.
Contaminated drinking water and unsafe sanitation practices contribute significantly to the high
incidence of diarrheal disease in rural Ethiopia (Usman et al. 2016). Repeated diarrhea episodes
and parasites infection also increase children’s risk of slow physical and mental development,
and poor school performance (Lorntz et al. 2006). The problem of child undernutrition and
substandard child growth attainment in the country is among the highest in the world. It is
estimated that 44 percent of under-five children are stunted, and 21 percent of children are
severely stunted (CSA & ICF International 2012). The proportion of child malnutrition is even
much higher in rural areas. For instance, the prevalence of child stunting (low height-for-age),
which often used as an indicator of chronic malnutrition), is over 50 percent in rural areas. As the
government had made swift measures to improve child malnutrition, the incidence of child
stunting has dramatically declined from 67 percent in 1992 to 44 percent in 2012 (UNICEF 2013).
Until these days, however, malnutrition and child mortality remain a great public health problem.
Table 1-1: Estimated drinking water and sanitation coverage 2015 update in percentage
Service types Year
1990 1995 2000 2005 2010 2015
Water supply Total improved 13 19 29 38 48 57
Piped onto premises 1 2 4 6 9 12
Other improved 12 17 25 32 39 45
Other unimproved 39 38 35 34 32 30
Surface water 48 43 36 28 20 13
Sanitation Improved 3 3 9 15 22 28
Shared 4 4 6 9 11 14
Other unimproved 1 2 9 16 23 29
Open defecation 92 91 76 60 44 29
Source: Author’s compilation using WHO/UNICEF JMP 2015 data.
3
It is understood that one of the health benefits from improved water supply happens through
better water quality—which reduces the ingestion of pathogens, which in turn reduces diarrhea
incidence and other intestinal parasites. On the other hand, providing means of safe excreta
disposal reduces the number of pathogens in the environment, which improves nutrient
absorption and diseases resistance. The provision of hygiene education, in addition to the
physical intervention, helps ensure that people preserve their health by washing hands with
soaps properly, removing feces safely, and store water more carefully (Bosch et al. 2002). It is
also argued that hygiene promotion and point-of-use (POU) water treatments are more effective
and sustainable to reduce diarrheal diseases; however, this long-held consensus is now
challenged by a recent review analysis by Waddington and Snilstveit (2009) and suggesting that
“sanitation ‘hardware’ interventions are highly effective and sustainable in reducing diarrheal
morbidity.” Provision of improved WATSAN services together with proper hygiene practices,
therefore, will significantly improve the country’s overall health, socioeconomic conditions, and
quality of life at large.
The government has made considerable efforts to improve the WATSAN situations of the country
by adopting the Universal Access Program (UAP) to provide access to safe drinking water for all
urban and rural population. To this end, it made major reforms by decentralizing service delivery
responsibilities to the woredas3 (districts) level and strengthening their capacity. The government
also adopts the Community-Led Total Sanitation (CLTS) approach to eradicating open defecation
practices in peri-urban and rural communities. The CLTS approach becomes highly promising in
influencing communal sanitation behaviors to stop open defecation, to construct their own pit
latrine, and to adopt proper hygiene practices. There are, however, significant challenges to
sustain the process and to keep up its momentum. Raising the level of awareness to change
people’s attitudes and practices are important for the CLTS to be effective over time. For
instance, if a minority in a given community continues to defecate in the open, everyone in the
community is at risk of infections. The government and the international organizations are
playing a key role in the CLTS in making a progress in the provision of WATSAN services in many
rural communities.
Agriculture is one of the primary sources of livelihoods for the vast majority of Ethiopian rural
households. Rain-fed subsistence agriculture dominates in the country but low and erratic rainfall
limits productivity and food security and resulting in an incidence of drought every 4-5 years
(Osman & Sauerborn 2002). Accordingly, small-scale irrigation (SSI) has been identified as a key
poverty reduction strategy because it is expected to increase productivity through the use of
available water resources. Irrigation is little developed and agricultural use of groundwater is low
in the country. Since access to improved water supply in rural areas is inadequate, separate water
sources for domestic purposes and agricultural uses do not exist in several areas. Consequently,
3 Woreda is the third lowest administrative unit in the country.
4
in such rural settings, drinking water supply and sanitation has complex interactions with
agriculture. Irrigated farming system is not well developed in the country, and such multiple-use
of water systems create competition for water between domestic and productive uses. This is
the central theme of this thesis, which makes it different from the pool of existing empirical
studies.
Although there is a trade-off in the domestic use of irrigation water, availability of irrigation water
may increase the amount of water for domestic uses. Where water is scarce, irrigation increases
water availability for domestic uses with the potential health benefits and saving time and
energy. On the other hand, using poor quality of irrigation water for domestic purposes may be
harmful to health. The threshold level of water quality which is acceptable for production
activities may not be acceptable for drinking or domestic uses. Moreover, the quality and
quantity of ground water for domestic uses may be affected by irrigation activities by reducing
the quality and availability of water within irrigated areas (Horgby & Larson 2013; van Der Hoek
et al. 1999). Irrigated agriculture may also increase food availability and dietary diversity to
enhance improved nutrition (Burney et al. 2013; Namara et al. 2005). In such a complex
environment, an integrated approach may be crucial to enhance improved WASH, efficient water
use, and sustainable food production. Improved health and nutrition may be ensured
simultaneously by exploring available opportunities to optimize the synergies and to reduce the
tradeoff between WATSAN and irrigated agriculture.
Considering the limited studies on the linkage between agriculture and WATSAN, this research
aims to shed more light on the agriculture-WATSAN nexus issues. Using a primary household
survey data, this study investigates the impacts of rural WATSAN services on child health
outcomes and the linkage between agriculture and WATSAN under multiple-use water systems.
The research findings bring a better understanding of the complex interactions between WASH
and irrigated agriculture under multiple-use water systems. This would provide necessary inputs
for policy makers and practitioners to design evidence-based policies/interventions regarding the
sector, which allows for effective investments and sustainable provision of improved WATSAN
and SSI infrastructures. It also serves as a guidance on how future investments should be
channeled to maximize the benefits of improved WATSAN services and small-scale agriculture in
support of improved health and nutrition outcomes.
1.2 Motivation and Problem Statement
Diseases caused by poor WATSAN services are major health problems in Ethiopia. It is understood
that most of the deadly diseases caused by poor WATSAN services are largely preventable. The
WATSAN problem is substantial in the country, and its WATSAN coverage is rated among the
lowest in the world—even by sub-Saharan African standards (WHO/UNICEF 2012). This shows
the need for improvement to avoid water-related diseases that cost many lives. It is estimated
5
that below 8 percent of the population in Ethiopia following the safe drinking of water chain from
source to mouth (Federal Ministry of Health (FMoH) 2011). In another report, the FMoH (2005)
indicates that 70 percent of the diarrheal diseases and 60 percent of the country’s disease burden
are mainly attributed to poor WASH. This suggests that inadequate WATSAN services and poor
hygiene practices cause thousands of the country’s poor people to die from preventable diseases
every year, which undoubtedly hampers the socioeconomic development of the country.
The FMoH estimates that the top 10 causes of mortality account for 74 percent of all deaths in
the country. Many Ethiopians face high morbidity and mortality with a growing prevalence of
communicable diseases mainly attributed to potentially preventable infectious diseases and
nutritional deficiencies. Although largely preventable, childhood and maternal illnesses and
communicable diseases are major causes of death in the country. Since access to improved
drinking water—both in terms of quality and quantity—is inadequate in rural areas, women and
children have to travel many hours to collect water far from home, and that may not even provide
safe water. Traveling long distance for water collection mean little is available for proper hygiene
because availability is limited. Such conditions and poor access to improved drinking water
sources make them more susceptible to the spread of infectious diseases.
Poor WATSAN services can affect the growth and development of young children. Acute and
chronic diarrhea has a significant impact on child growth outcomes. Repeated diarrhea episodes
sap the nutritional status of children and reduce the availability of nutrients by decreasing the
body’s capacity to absorb nutrients which led to stunting (reduced child growth rate), impaired
physical growth, and cognitive development (UNICEF 2013; Humphrey 2009). Moreover,
malnourished children have lower resistance to diseases, and they are more likely to suffer from
subsequent diarrhea and other infections (UNICEF 2013). For instance, severely undernourished
child is almost 10 times higher than average to die from diarrhea (Black et al. 2008). This vicious
cycle has a strong negative influence on child growth and development.
To address the problem of access to improved WATSAN services in the country, the sector got an
eye from the government of Ethiopia (GoE) and international development partners. They have
been working to increase access to WATSAN services at household, community and in many
public areas, and to bring about basic hygiene behavioral changes; including hand washing with
soap at critical times, and safe water handling and treatment. Despite all this progress, many
Ethiopians still lack clean drinking water and basic sanitation facilities. The widespread open
defecation practices and frequent outbreaks of diseases due to inadequate WASH services
remain to be major challenges of the country. To substantially reduce the spread of
communicable diseases in the country, still much work needs to be done in the WASH sector.
The provision of improved WATSAN services and proper hygiene practices are of crucial
importance in the prevention of waterborne diseases, including diarrhea (Usman et al. 2016). It
is estimated that improved water supply can result in an average reduction in cases of diarrheal
6
morbidity over 24 percent; improved sanitation facility reduces by 33 percent; and proper
hygiene practices, such as handwashing with soap at critical times, have been found to reduce
diarrheal incidence by more than 40 percent (WHO/UNICEF 2009). Investments in WATSAN
infrastructures generally result in higher economic returns, which contribute to overall economic
growth—particularly in developing countries. Depending on the region, US$1 invested in
WATSAN would give an economic return of between US$3 and US$34 (Hutton & Haller 2004). It
is also argued that “people without access to water supply and sanitation are […] caught in a
vicious cycle in which lack of access leads to poverty which in turn prevents people from having
the ability to gain access to services. Lack of access to water and sanitation is thus both a cause
and an effect of inequality, poverty, and underdevelopment” (Brocklehurst 2011, p.6).
On the other hand, in rural Ethiopia, SSI has a complex interaction with rural drinking water
supply, which is far below the desired services levels (Butterworth et al. 2013; Scheelbeek 2005).
Irrigated agriculture and WASH services are interconnected, and they can have different
implication on health and nutrition outcomes. The following Figure 1-1 shows the conceptual
linkage between agriculture, nutrition, and health. This conceptual framework is discussed in
detail in section 5.2. Irrigation can impact health and nutrition outcomes through several
pathways. It can influence health negatively through increased malaria incidence. For instance,
certain agricultural development practices aimed at increasing productivity to ensure food
security, such as water harvesting techniques, irrigation canals, ponds, tanks and/or dams, can
actually exacerbate the incidence of diseases and create suitable conditions for the propagations
of waterborne diseases-vectors such as mosquitoes transmitting malaria (Amacher et al. 2004;
Asayehegn 2012; Asenso-Okyere et al. 2012; Ersado 2005; Hawkes & Ruel 2006; Keiser et al.
2005; Kibret et al. 2010; Kibret et al. 2009). Irrigated agriculture also influences health through
other channels. Irrigation often increases household’s income which allows them to access
improved health care service. In addition, domestic water quality and quantity might be affected
by irrigation practice. Irrigation water is often used for non-agricultural purposes, including
domestic uses such as for food preparation, gardening, laundering and personal hygiene.
Irrigation water can also be the sources of drinking water in developing countries where
improved drinking water supply is inadequate (van Der Hoek et al. 2001; van Der Hoek et al.
1999). Although, in most cases, the level of water quality for irrigation may not be acceptable for
domestic consumption, van der Hoek et al. (2002) argue that water quantity is more important
than water quality as it reduces the prevalence of hygiene related diseases.
Although there is a growing understanding in the literature that agriculture, nutrition, and
health—especially through WATSAN—are strongly linked, determining those linkages in their
dynamic, trade-offs and interactions between them remain to be a challenge for policy-makers
and development practitioners and has to be examined properly. Any development projects in
developing countries that attempt to address agricultural and nutritional issues are less likely to
7
meet their targets unless the role of WATSAN services in improving health and nutritional
outcomes have been recognized and properly addressed.
Figure 1-1: The linkage between agriculture and nutrition and health
Source: Author’s own illustration.
1.3 Objectives
Given the discussion in the previous sections, this thesis addresses the research question, “how
water, sanitation, and agriculture are linked, and what are their implication for health and nutritional
outcomes in rural Ethiopia?” More specifically, this study has the following four specific research
objectives:
(i) To examine the key drivers of household drinking water quality in rural areas.
This objective examines the primary factors that influence household storage water quality. Here,
we tried to investigate the influence of socio-demographic, types of water sources, water
collection vessels, sanitary and community characteristics on stored household drinking water
quality. The outcome variable is the count of E.coli colony-coliform units per 100ml (CFU/100ml)
water, and a binary variable based on contaminated and uncontaminated storage water quality.
8
(ii) To assess the impacts of drinking water quality and sanitation practice on childhood
diarrhea in the study areas.
This objective investigates the impacts of quality of stored drinking water and household
sanitation and hygiene behavior on diarrhea prevalence among under-five children in the study
districts. We used instrumental variable approach to address this research question. In addition,
it examines the longitudinal diarrhea prevalence on under-five children using longitudinal data.
We followed under-five children of the same households biweekly visiting them for a period of
12 weeks to record the prevalence of diarrheal and other infectious diseases.
(iii) To study the linkages between rural water supply, sanitation, and agriculture, and
their impacts on health outcomes.
This is the principal theme of this thesis and focuses on the linkage between irrigated agriculture
and WASH, as well as their impacts on household’s health outcomes and storage water quality.
Health was measured as the prevalence of diarrhea (diarrhea and cholera diseases symptoms),
water-related diseases (skin-itching, typhoid fever, giardiasis, and trachoma), and overall
morbidity status (excluding injuries, chronic and other unknown diseases symptoms) in the
preceding two months before the survey. In this section, we primarily focus on non-preschool
household members. The analysis is based on both individual and household level information.
In the empirical analysis, we focused on diarrheal diseases, but interested readers can find the
estimation results for the other health outcome measures in Appendix A.
(iv) To investigate the association between households water quality and sanitation
practices and children’s growth outcomes.
The nutrition outcome analysis focuses on underweight and stunting for under-five children. We
use anthropometric measures of weight and height to calculate weight-for-age, height-for-age,
and weight-for-height -z-scores to determine whether a child is underweight, stunted and/or
wasted, respectively, using the new WHO 2006 reference group.
1.4 Significance of the Study
The findings of the research, in addition to filling the research gap in the empirical literature, it
would provide key inputs for policy makers and development practitioners to design evidence-
based WATSAN interventions and to make informed decisions. Such interventions would likely
be effective and sustainable to ensure quality provision of WATSAN services through the full
engagement of all relevant actors. Building the capacities of rural communities to maintain and
manage their WATSAN environment by themselves can be instrumental in improving their well-
being and lives at large. It also helps to identify investment priorities at the interface of
agriculture-WATSAN nexus for better health and nutrition outcomes for key players of the sector
such as the regional government, local and international developmental partners. Unlike many
9
existing empirical studies in this area, this study is a first-attempt to holistically investigating the
linkage among agriculture, WATSAN, and health using econometric methods. Mostly, existing
empirical works focus on the relationship either between WATSAN and health or irrigation and
health while isolating the complex interaction between agriculture and WATSAN in rural settings.
Furthermore, due to its interdisciplinary nature, this research is in the frontiers of the ZEF
research themes, and it could provide some insights for future research on many development
aspects such as water resources management, reducing extreme poverty and hunger, promoting
gender equality and empowerment, improving maternal health, and reducing child morbidity
and mortality through combating waterborne diseases.
1.5 Organization of the Thesis
This thesis has seven chapters. Chapter 1 provides a general background information, and it
briefly discusses the problems, research objectives and its contribution to the wider empirical
literature. Following the preceding introductory discussions, chapter 2 provides a detail
description of the study areas, data collection process, the design of the household survey, and
the water and anthropometric measurements during the field work. Chapter 3 presents an
empirical study on the key drivers of household storage water quality. It examines the primary
factors affecting the microbial quality (Escherichia coli bacteria (CFU/100ml water sample)) of
stored household drinking water. It employs simple ordinary least squares (OLS) and logistic
model estimation. Chapter 4 investigates the impacts of storage water quality and household
sanitation practices on diarrhea among under-five children. The prevalence of diarrhea is a self-
reported prevalence by the primary caretaker of the child in the preceding two weeks before the
survey. Methodologically, an instrumental variable (two-stage least square) estimator and
recursive bivariate probit methods have been used. In this chapter, we briefly examine the
longitudinal prevalence of diarrhea using a longitudinal data collected from the same households
during the follow-up survey for a period of 12 weeks. We also examine a cost-benefit analysis for
an improved water supply in the study areas.
Chapter 5 addresses the research question on the linkage between domestic water supply and
irrigation and their implications on health outcomes and domestic water quality/quantity. This
chapter exclusively focuses on non-preschool household members. Although diarrhea is a
measure of our primary interest outcome variable, other health outcome indicators, such as
other water-related and general morbidity status, have also been investigated, particularly in the
descriptive analysis section. Also, the unit of analysis focuses on both household and individual
level for a more robust relationship. Chapter 6 studies the determinants of child growth outcome.
Based on the anthropometric measures of weight and height, it primarily focuses on underweight
and stunting. Finally, chapter 7 presents the major findings of the study, with potential policy
implications and suggestions for further research directions.
10
2. Data Collection and Survey Descriptions
2.1 Description of the Study Areas
The overarching objective of this study is to investigate the complex interactions and linkage
between irrigated agriculture and WASH to provide a set of evidence-based recommendations.
In doing so, it identifies investment priorities and policies to improve the WATSAN environment
of rural communities, which in turn improves health and nutrition outcomes. It is conducted as
part of a broader thematic area of “guiding pro-poor investments in the nexus among ‘domestic
water quality and quantity’, ‘sanitation and hygiene’ and agriculture from the bottom-up (AG-
WATSAN Nexus).” The project has been undertaking in Ethiopia, Ghana, India, and Bangladesh
where access to WASH is seriously lacking. Considering the central theme of the project, the study
sites were identified based on ex-ante information on access to improved WATSAN, the
prevalence of waterborne and water-washed diseases, and the practice of irrigated agriculture
by taking into account the multiple-use of water resources. In addition, different stakeholders in
government and non-government organization (NGOs), which are closely working in the WASH
sector, were consulted during the hotspot mapping process.
To identify suitable areas for this study, the above ex-ante information was factored in the
hotspot mapping process. Following that line, five districts in the Amhara and Oromia regions
were identified and two of them were visited. Discussion with experts from government and
NGOs working in the sector were made to learn more on the suitability of identified districts.
Succeeding survey field assessment, Fogera and Mecha districts were identified because existing
coverage of improved WATSAN infrastructure is among the lowest, the prevalence of waterborne
and water-washed diseases are also primary health problems of the inhabitants, and SSI adoption
is common. Figure 2-1 depicts the map of the study areas. The shaded area with light-green
represents the Amhara National Regional State (ANRS) map, and the two selected districts
(shaded in blue) are shown in the right panel of the figure.
For the reasons discussed earlier, the study was carried out in Fogera and Mecha districts, and
these areas are more appropriate in terms of providing the platform to investigate the linkages
and interactions between irrigation and domestic water supply and to identify possible channels
through which AG−WATSAN nexus works and results in better health and nutritional status. To
reduce the burden of conducting complex household and community level surveys, logistic and
institutional factors were also taken into account in the selection of these districts.
11
Figure 2-1: Map of the study areas
Source: Author’s own illustration.
Fogera woreda
Fogera district is one of the 106 woredas of the ANRS and found in South Gondar Zone. It is
situated at 110o 58o N latitude and 37o 41o E longitude. Woreta is the administrative town of the
district and is situated 625 km north of Addis Ababa and 55 km from Bahir Dar—the capital city
of the ANRS. The district is divided into 29 rural and 5 urban Kebeles.4
In terms of geography, the district consists of plain or flat lands (which accounts for 76%),
hills/mountains (11%) and 13 percent valley bottoms classified locally as highland, midland, and
lowland topography levels. Its altitude ranges from 1,774 up to 2,410 meters above sea level,
which allows a favorable opportunity for a range of crop production and livestock rearing (ILRI
2005). The mean annual rainfall is 1216.3 mm, with Belg (short rainy season) and Meher (long
rainy season) cropping seasons.
The total land area of the district is 117,414 ha, of which about 43 percent is cultivated for the
production of various annual crops. Most of the farm land was allocated for annual crops where
cereals covered 51,472 ha; pulses cover 9,819.98 ha; oil seeds 6,137 ha; root crops 1,034.29 ha;
and vegetables 882.08 ha (CSA 2003). In order of area coverage, the major crops include teff,
4 Kebele is the lowest administrative unit in Ethiopia.
12
maize, finger millet and rice. As stated by the ILRI (2005) report, average land holding size was
about 1.4 ha with minimum and maximum of 0.5 and 3.0 ha, respectively.
In addition to rain-fed subsistence agriculture, most farmers practice irrigation farming using
various water sources. Farmers use rivers, such as the Gumara and the Ribb, which drain into
Lake Tana, as a primary source of water for irrigation activities. River diversion is the predominant
system of irrigation methods, but some farmers also use motor pump technology for irrigation.
Most farmers practice traditional irrigation systems mainly to produce onions and tomatoes.
Mecha woreda
Mecha is the other district where the study is conducted, and it belongs to the west Gojjam Zone
in ANRS. Merawi is the administrative town of the district and is located about 525 km North of
Addis Ababa and 34 km southeast of Bahir Dar. Its climatic condition alternates between a long
summer rainfall (June—September) and winter dry season (December—March) with mean
annual rainfall of 1200-2000 mm. The mean temperature varies between 24—27oC, and its
altitude range from 1800 to 2500 meter above sea level. The study area is located at latitude
10o 30o N and longitude 37o29o E.
Agriculture is the key economic activities in the study area, employing nearly 100 percent of labor
forces. Mixed farming is the dominant agricultural practice. Smallholders mainly produce both
staples crops—typically wheat, barley and teff, and cash crops—such as pepper and potatoes.
The Koga irrigation watershed is the principal source for irrigation activities in the area. The Koga
irrigation project benefits more than 14,000 households and helps boost agricultural production
in the area around Merawi, south of Lake Tana in the ANRS. As of July 2012, the population of
Mecha and Fogera district is estimated to be 334, 789 (with an area of 1,481.64 sq. km) and 264,
512 (with an area of 1,111.43 sq. km), respectively (CSA 2013).
WASH and prevalence of WASH-related diseases in the study areas
The provision of improved WATSAN in rural Ethiopia is far behind the desired service levels. For
instance, access to safe drinking water in Fogera district is estimated to be 69.9 percent (Stel &
Abate 2014); while another report estimated even a much lower figure (42%). Furthermore,
some community water sources failed to operate regularly due to poor maintenance. On the
other hand, only 35 percent of the rural population of Mecha district have access to improved
water sources (Beyene 2012). Furthermore, most water sources do not provide sufficient water
and become shrink during the summer. Consequently, most rural households depend on
unimproved sources, such as unprotected wells/springs, rivers, which are easily polluted by
human and animal wastes. The coverage of simple pit latrine in the study areas is low. In some
cases, households with a pit latrine do not use it frequently and prefer to defecate in the open.
The government together with local and international development partners are striving to help
rural communities to improve access to improved WATSAN services.
13
Increasing access to improved WATSAN services is key development goals for the Ethiopian
government. Until these days, however, more than half of the rural households do not have
access to improved water sources. Moreover, basic sanitation and disposal facilities that separate
waste from human contacts are extremely low, which can seriously affect the health status of
the communities. Accurate data on the actual coverage of sanitation and key hygiene indicators are
not available at the district level. On the word of the 2013 Mecha health center report, pneumonia,
diarrhea (non-bloody), malaria PF, acute febrile illness (AFI), acute upper respiratory infection,
malaria PV, and skin infections are reported to be the primary health problems for under-five
children. Since the coverage of WATSAN and key hygiene practices are lacking, the incidence of
waterborne and water-related diseases are serious public health problems.
2.2 Data Collection and Organization of the Survey
The AG-WATSAN nexus project had been undertaken in collaboration with the Ethiopian
Economics Association (EEA) through its research wing called ‘Ethiopian Economic Policy
Research Institute’ (EEPRI), Addis Ababa, Ethiopia. The household survey was carried out under
the support of the EEA and was implemented by the Center for Development Research (ZEF),
University of Bonn, Germany. The resources and materials for conducting the survey were
provided by the regional office of Central Statistical Agency (CSA) of Ethiopia, Bahir Dar branch,
Organization for Rehabilitation and Development in Amhara (ORDA), German Agro Action
(Welthungerhilfe), and health center of the selected districts.
a) Training, pretest, and fieldwork
Following recruitment of field interviewers and supervisors for the field work in Bahir Dar,
interviewers and supervisors trained during the first week of February 2014. Since the quality of
data obtained from a survey highly depends on data collectors and supervisors, the necessary
training was given to the teams to equip them with skills and techniques required for the survey.
The training includes instruction on interviewing techniques and survey procedures, detailed
review of the questionnaires contents, instructions and practices in measuring height and weight
of children, global positioning system (GPS) reading, and mock interviews among themselves.
The Amharic language questionnaires were primarily used during the training, while the English
versions were simultaneously checked against the Amharic questionnaires to ensure an accurate
translation.
Before the baseline survey, the questionnaires were pretested in Fogera district, Kuahar Michael
kebele, to make sure that the questionnaires were clear and can be understood by potential
respondents. Anthropometric measurements and water sample quality testing were also
conducted during the pretest field work. Finally, the questionnaires were revised based on
lessons drawn from the pretest exercises.
14
Water sample quality-testing and child anthropometric measurements were performed
simultaneously during the baseline survey. The microbiological quality of stored household
drinking water and primary community water sources were tested for fecal contamination.
Escherichia coli (E.coli) bacteria (CFU/100ml water) is used as an indicator as a proxy for fecal
contamination, which is considered to be a major cause of diarrhea. The survey also employed a
GPS reading to identify each selected households.
b) Fieldwork and data processing
The survey was carried out by two interviewing teams. Each team consisted of a supervisor (a
supervisor is assigned to closely manage and follow-up the work of five enumerators), a water
sample collector and health professionals (health extension workers). On top of that, a water
sample analyst also participated in the field work. To improve the quality of collected data, most
of the questionnaires were checked by both the field supervisors and other external editors. A
baseline survey was administered between February and March 2014. After the completion of
the baseline survey, all questionnaires were returned to Addis Ababa for data processing, which
consisted of office editing, coding of open-ended questions and data entry. The data were
processed by a team of four data coders and editors, two data entry operators as well as a data
entry supervisor. The data entry and editing processing were started in March and completed in
July 2014.
2.3 Survey Design, Sampling Frame and Sample Size
Fogera and Mecha districts were purposely identified from the ANRS based on access to
improved drinking water and sanitation coverage, the prevalence of waterborne and water-
related diseases and availability of small-scale irrigation practices. The next step in the sampling
process was to select sample households from these districts.5 Administratively, regions in
Ethiopia are divided into zones, which are subdivided into administrative units called woreda
(district). Each district is further subdivided into the lowest administrative unit, called kebele. To
select the sample agricultural households, a total of 20 kebeles were identified based on multiple-
use water systems and accessibility: 11 kebeles (8 irrigating and 3 non-irrigating) from Fogera and
9 kebeles (6 irrigating and 3 non-irrigating) from Mecha district.
The sampled households were selected using a stratified two-stage cluster sample design,
whereby villages were taken to be the primary sampling units (PSUs), and agricultural households
were considered to be the secondary sampling units (SSUs). The first-stage sample selection of
villages/clusters was random but proportional to size (PPS). Subsequently, the second-stage
sampling units, agricultural households, were selected within each village based on the
systematic random sampling (SRS) method. The lowest geographical administrative division of
5 A household is a group of people who live together and take food from the same plate, and someone who has lived
in the household at least six months.
15
the region (i.e., kebele) is used to form the first level of stratification. In the first stage of the
sample selection process, 61 clusters/villages were randomly selected from 20 administrative
kebeles. This was followed by a systematic random sample selection of 454 households from a
complete listing of agricultural households. This implies that within each stratum the villages are
used as PSUs and agricultural households as SSUs. The household head lists were provided by the
Woreda’s Agriculture and Extension Bureau. Among the selected households, 302 households
engaged in irrigated agriculture and the remaining 152 households do not.
One of the main selection criteria for a household was having at least one child under-five years.
As such data were not available beforehand—during the first stage selection process, households
without under-five child were replaced by the nearest neighbor households during the fieldwork.
The overall sample size was 454 households with 565 children under-five years old. The sample
size was determined by taking into account both the required level of precision for the most
important indicators and the amount of time and resources available to the survey.
2.3.1 Questionnaires and ethical consideration
Survey questionnaires
The AG-WATSAN household survey collected a range of information using structured module
questionnaires:6 household, WATSAN, primary caregiver, child and community questionnaires.
Most of these questionnaires were adapted from various standardized national survey
instruments. In addition to the English language, the questionnaires were translated into
Amharic—the major language of the study area.
The Household questionnaire was used to list all the members of the selected households and
basic characteristics of each listed person, such as age, sex, education, relationship to the
household head and other household level information. The household questionnaire also
collected information about total agricultural productions and sales, livestock ownership,
holdings of various consumer and durable goods and other income sources, food and non-food
expenditures, labor and time use, and household members’ health status in the last two months
preceding the survey. The WATSAN module was used to collect information on characteristics of
housing, water supply sources, continuity of water supply and seasonal change, water storage,
handling and treatment, community management, toilet facility, waste management, and
irrigated agriculture practices. The Primary caregiver questionnaire was used to collect
information about hygiene behavior and knowledge, community participation, handwashing
demonstration, and food preparation, storage and handling. Furthermore, the Child
questionnaire was used to collect information about child feeding practices such as exclusive
breastfeeding during the first six months of life, current breastfeeding and complementary foods,
6 The various questionnaire modules are available at: https://data.zef.de/?uuid=4ac3e43a-557e-4ddc-8f58-
d2b51404c581
16
prevalence of illness in the previous two weeks period (such as diarrhea, malaria, fever,
pneumonia, and cough), immunization records and anthropometric measurements. In most
cases, the respondents were the primary caretakers concerning household consumption,
WATSAN and child health. Moreover, GPS readings were used to identify the relative position of
sampled housing units.
Ethical consideration
The study protocol was approved by the Medical Ethic Commission of the Federal Republic of
Germany (Arbeitskreis Medizinischer Ethikkommissionen in der Bundesrepublik Deutschland).
Community leaders and household’s head provided informed consents. During the field work, a
two-page consent form was read out to a household head with an introduction and purpose of
the research, types of research intervention, voluntary participation, participation benefits,
confidentiality, and right to refuse or withdraw from the study at any stage. Two copies of the
consent forms were prepared, and one copy was given to the participant household (a copy is
attached in Appendix B).
2.3.2 Anthropometric measures and water sample testing
a) Height and weight measurements
The AG-WATSAN survey data included height and weight measurements. During the fieldwork,
interviewing teams obtained measurements of height and weight of all under-five children listed
in the household roster. The primary caretaker’s willingness consented to the measurements.
The interviewing team was into two teams, and each team carried a balance scale and measuring
board. The scales were electronic scales with a digital screen. Height measurements were carried
out using measuring boards made of wood. The scales and measuring boards were obtained from
the regional office of Ethiopian CSA, Bahir Dar branch office. These resources are mainly designed
for use for household survey settings. For instance, they were used by Ethiopian Demographic
and Health Survey (EDHS). The measurements were carried out by the local health extension
workers (HEWs). Following the DHS standards, children younger than 24 months were measured
lying down on the board (recumbent length) and older children were measured while standing.
The weight of young children was obtained by subtracting his/her mother’s weight from the
combined weight of mother and child. Weight is measured to the nearest 100 grams while height
is measured to the nearest 0.1 cm.
b) Drinking water sample quality-testing
In addition to collecting information on how households handle and store their drinking water,
the quality of the water they drink was tested in each selected household. Drinking water samples
were collected for testing the microbiological level of the water to determine the degree of its
health risks. We used a Potatest Incubator Kit to test water samples in the field. The Potatest
detects the presence of E.coli bacteria which indicates the level of fecal contamination of the
17
water. Microbiological water quality tests were also performed for a total of 61 water samples,
which were taken directly from community water sources in each selected villages.
The test kit was provided by the German Agro Action (Welthungerhilfe)—an NGO working in the
WASH sector in rural and peri-urban areas of the country. Enumerators asked household
members—usually an adult woman—the following question: ‘could you please give me some
water for drinking?’ so that their behavior would not be altered. Then the water samples were
kept in coded bottles, which were properly sterilized using autoclaves from the local health
centers. The collected water samples immediately placed into the portable test kit on-site and
incubated for a maximum of 24 hours at the temperature of 44 degree Celsius. The bacteria
colonies grew on the Membrane Lauryl Sulfate Broth (MLSB) medium, which is specifically
formulated to facilitate the growth of E.coli bacteria and prevent the growth of other micro-
organisms. Since bacteria are extremely small, they should be grown on nutrient plates so that
they can multiply rapidly and, depending on the test kit, become visible within 24 hours. Upon
completion of the incubation period, we enumerated the number of E.coli concentration, which
is reported as colony-forming units per 100ml (CFU/100ml) water sample (see section 3.3 for
further description of the water sample quality-testing procedure).
2.3.3 Follow-up survey
The follow-up surveys were administered between April and July 2014 after completing the
baseline survey by the end of March 2014. The aim of the survey was to collect information on
under-five children health status—such as diarrhea, fever, vomiting, constant cough, stomach
pain/cramps and skin infections in the preceding two weeks. This data collection process requires
multiple successive visits to the households, and each household has been visited every
fortnightly for a period of three months to record the health status of the children. For this
purpose, one-page child health calendar questionnaire was developed to record information
about symptoms of a particular illness; how long the symptoms stay; and what treatment they
seek for if any. The data collection was done by health and agriculture extension workers who
work in the selected kebeles. Training was given to the enumerators during the field organization
for the follow-up household survey. In most cases, the data collectors ask the child’s primary
caretaker—usually the mother or another adult woman household member.
18
3. Drivers of Microbial Quality of Household Drinking Water in Rural Households
3.1 Introduction
Lack of access to safe and adequate water supply and the health risks associated with water-
related diseases are major public health problems in many developing countries. Today, more
than 700 million people, who mostly live in the developing countries, are without access to
improved and adequate water (WHO/UNICEF 2014). More than 1.5 million under-five children
die of diarrheal diseases every year (WHO/UNICEF 2009). Unsafe drinking water is considered to
be one of the major causes of diarrhea (Zwane & Kremer 2007). Increasing the provision of
improved drinking water is crucial in the fighting against diarrheal diseases for young children in
developing countries.
Figure 3-1 presents the coverage of access to improved water supply in Ethiopia. About 57
percent of households have access to an improved drinking water source, with a higher
proportion of urban residents (93%) than among rural residents (49%). The disparity between
urban and rural households is big in terms of access and types of services. Moreover, as access
to improved sanitation facilities are inadequate in rural areas, the majority of households
defecate in the bush or open fields (WHO/UNICEF 2015). Furthermore, including drinking water
safety or quality criteria in the WHO/UNICEF JMP definition of access to improved drinking water,
the reported figures would be substantially lower in both urban and rural Ethiopia, because water
collected from improved sources are often re-contaminated during collection, transportation,
and storage (Wright et al. 2004). Consequently, the current definition used by the JMP is likely to
lead to substantial overestimation of the number of population who have access to improved
water sources in many developing countries (Bain et al. 2014; Godfrey et al. 2011). In rural
Ethiopia, hand-pump water sources are also often broken and non-functional due to poor
maintenance and repairs (MoWE 2007). This would further reduce the actual number of
households reported to have access to improved water sources.
Ethiopia is considered as the water tower of East Africa, yet in most parts of the country, water
is still as inaccessible as it is precious. Moreover, water quality is poor and often contaminated
by human and animal feces. As a result of limited improved water availability, most rural
population relies on unimproved water sources. People use unprotected springs, shallow wells,
irrigation water from canals and rivers as a source of water for domestic uses which are easily
polluted by human and animal feces. Unimproved sanitation habits and open defecation
practices exacerbate the problem. Often shallow and unprotected community water source
points are subject to gross contamination when rainwater washes wastes from surrounding areas
into the sources. The situation is much worse where drinking water sources are shared with
livestock. During the dry season, most of the traditional water sources are placed under pressure
as shallow wells, springs and several other recurrent sources of water shrink-up. Moreover, due
19
to long distance and queues, rural households can only collect few liters of water for drinking and
cooking (Sutton et al. 2012). Limited availability of water may also prevent basic personal hygiene
practices. Such conditions are believed to contribute to high rates of morbidity and mortality
particularly in rural areas of the country (Begashaw 2003).
Figure 3-1: Percentage of population by type of drinking water source in Ethiopia, 2015
Source: Author’s compilation using the WHO/UNICEF JMP dataset.
There are various chemical, physiological and microbiological standards for a water supply to be
qualified and acceptable for drinking. While water contamination can have various origins, this
study primarily focuses on E.coli bacteria—one of the most common indicators for microbial
water quality studies. This bacteria comes only from human and animal excreta. Human feces
are the primary source of pathogens that cause waterborne diseases such as diarrhea. As per the
WHO drinking water quality guideline, E.coli bacteria as a microbial water quality indicator should
be zero per 100ml for the water to be considered safe for drinking (WHO/UNICEF 2010a). A single
gram of human feces can contain 10 million viruses, one million bacteria, 1,000 parasite cysts and
100 parasite eggs (UNICEF 2000). These pathogens can transfer from an infected host to a new
one via various routes. They can also easily get into water supply sources where sanitation
facilities are inadequate and open defecation is widespread. In rural areas, water source
contamination is more pronounced because most water supply sources are inadequately
protected (Butterworth et al. 2013) and improved latrines are limited. However, removing
human excreta safely and cleaning hands with soaps after contact with fecal material
substantially reduces the transmission of pathogen agents (Curtis & Cairncross 2003).
56
112
37
4845
6
3530
1
1613
0
10
20
30
40
50
60
Urban Rural Total
Per
cen
tage
of
po
pu
lati
on
Piped into premises Other improved sources Other unimproved sources Surface water
20
Determining the public health risk associated with drinking water quality is important. To
determine the health risk, the WHO recommends a routine monitoring of drinking water quality
but this is generally not feasible in the context of rural Ethiopia because either the analytical tools
often do not exist or the tests are expensive and complicated to perform (WHO/UNICEF 2010a).
Given the complexity of water quality problem at the POU, subjective judgments about storage
water quality based on the types of sources are often misleading in the absence of household
intervention to improve water quality at the POU.
This chapter aims to identify the factors that influence the quality of drinking water stored in the
households7 in two rural districts of Ethiopia. It investigates the quality of storage drinking water
and community water sources at a large scale in multiple-use water systems of rural Ethiopia
where improved WATSAN infrastructures are limited. This chapter has two major contributions.
First, existing studies that examine the determinants of storage water quality and its relationship
with rural water supply sources and household’s sanitary behaviors are quite limited: they
primarily focus on the impact of water source types on storage water quality and ignore hygiene-
and sanitation-related factors (Amenu et al. 2014; Yasin et al. 2015). Second, determinants of
domestic water quality under multiple-use water systems is understudied (Scheelbeek 2005;
Sutton et al. 2011). Irrigated agriculture has a complex interaction with domestic water in rural
areas as irrigation water often uses for a range of activities such as drinking, cooking, and bathing
among others. This type of work is, therefore, crucial to enhance the understanding of the
determinants of the microbial quality of storage water in rural households of Ethiopia and might
thus help policy makers to design the right intervention to improve access to safe drinking water
in rural areas.
3.2 Context and Related Literature
Ethiopia has made remarkable progress to improve the WATSAN situations of the country by
adopting the Universal Access Plan (UAP) in 2005. It aims to provide access to safe drinking water
for all rural and urban population of the country before the end of 2015 (MoWE 2006). This was
an ambitious target to be realized. Ethiopia’s UAP defines the minimum standards for rural
population as at least 15 liters of water for every one per day within 1.5 km of their home.
Although the government is playing a key role in the rural water supply schemes, the role of NGOs
and its development partners have been crucial since the government does not have the financial
resources and/or the technical capacity to undertake this radical and ambitious move alone.
To increase access to safe drinking water in rural areas and to provide 15 liters of water per day
for everyone within 1.5 km radius, several on-spot springs protection, normal hand dug wells,
and hand dug wells with pump ropes have been constructed in many rural areas (MoWE 2006).
As most of these water supply points fail to function just after their installation, sustainability
7 Hereafter storage water.
21
issues become a major challenge in the provision of safe water supply in rural areas. For instance,
a survey of water source points in rural Ethiopia found that 29 percent of hand-pumps and one-
third of mechanized boreholes were not functioning mainly because of maintenance and repair
problems (UNDP 2006). The 2012 National Water Inventory (NWI) report also indicates that more
than 93,000 water schemes across the country were non-functional. Moreover, studies in
Ethiopia indicate a strong relationship between rural water supply functionality and governance
(Welle & Williams 2014). To make the matter worse, most existing community water sources are
often contaminated with fecal materials and pose a high public health risk (Amenu et al. 2014;
Atnafu 2006; Jano 2007; Tsega et al. 2014).
The WHO/UNICEF JMP for WATSAN defines access to drinking water and sanitation in terms of
the types of technology and levels of service provided. The WHO sets five basic indicators for a
safe water supply such as water quality, quantity, cost or affordability, continuity, and coverage
or accessibility. Table 3-1 shows the current WHO/UNICEF JMP classification of improved or
unimproved WATSAN technologies. This definition of access to ‘improved’ water source,
however, does not consider the safety or quality of the water; subsequently, it does reliably
predict neither the microbiological nor the physiological quality of the water being consumed. As
this approach can be highly misleading, it is argued that inclusion of water safety parameter will
further reduce the coverage level of improved water sources reported by JMP due to the high
risk of microbiological contamination (Bain et al. 2014; Godfrey et al. 2011).
Table 3-1: JMP Classification of drinking water source types and sanitation facilities
Category Types drinking-waters sources Types of sanitation facilities
Improved Piped water into dwelling, yard/plot, Public
tap/standpipe, Tube-well/borehole, Protected dug
wells, Protected spring, and Rainwater collection
Flush/pour-flush to piped sewer
system/septic tank/pit latrine, ventilated
improved pit (VIP) latrine, pit latrine with
slab and composting toilet
Unimproved
Unprotected dug wells, Unprotected spring, Cart
with small tank/drum, Tanker truck-provided water a, Surface water (river, dam, lake, pond, stream,
canal, irrigation channel) and Bottles water b
Flush/pour-flush to elsewhere (that is not
piped sewer system, septic tank or pit
latrine), pit latrine without slab/open pit,
bucket, hanging toilet/hanging latrine,
shared facilities of any type and no
facilities, bush/field a Normally considered being “unimproved” because of concerns about the quantity of supplied water. b Considered to be “unimproved” because of concerns about access to adequate amount of water, about inadequate treatment, or about transportation of the water in inappropriate containers.
Source: WHO/UNICEF (2010b).
The WHO/UNICEF report presented in Table 3-2 provides some evidence on the status of
microbial water quality in Ethiopia at the national level. The result shows that, of the 1602 water
samples analyzed for thermotolerant coliforms (TTC), 1153 of 1602 (72%) samples met both the
22
national standard and the WHO guideline value of <1 CFU/100ml water. However, 7 percent had
counts of 1−10 CFU/100ml water, and another 14 percent had counts of 11−100 CFU/100ml
water. Overall, 7 percent of all samples had counts >100 CFU/100ml water. The proportion of
11−100 CFU/100ml and >100 CFU/100ml water count is significantly higher for protected springs
and protected dug wells but it is lower for utility piped supplies because they are better protected
than other water source points. Utility piped supplies are also often chlorinated which protects
the water from microbial contamination (WHO/UNICEF 2010a).
Table 3-2: Compliance of drinking water sources in Ethiopia for thermotolerant coliforms a
Count category
(CFU/100ml)
Utility piped supplies
Boreholes Protected springs Protected dug
wells Total
Prop. (%) Prop. (%) Prop. (%) Prop. (%) Prop. (%)
<1 87.7 67.9 43.3 54.8 72.0
1-10 4.2 9.9 10.0 11.0 6.9
11-100 6.4 16.9 29.2 21.3 14.3
>100 1.8 6.2 17.6 12.9 6.8
Sources sampled (n) 838 290 319 155 1 602
a CFU=colony-forming unit. Prop. = proportion of water samples showing corresponding count category.
Source: Adapted from WHO/UNICEF (2010a, p.21).
Few studies in Ethiopia examine the chemical and microbial quality of drinking water. Existing
studies related the water quality aspects with seasonality, type of water sources, and storage
behavior. Amenu et al. (2014) investigated the microbial water quality of rural households in
Lemu and Siraro districts of Oromia region. A total of 233 water samples collected from
household’s drinking water (126 collected during dry and 107 samples collected in wet seasons)
were analyzed. The study finds that about 55 percent of the samples were contaminated with
E.coli; however, the concentration of E.coli was much higher during the wet season than the dry
season.
The issue of seasonality and water quality was addressed in other studies. For instance, Sandiford
et al. (1989) examined 150 water samples from rural Nicaragua for the fecal coliforms
contamination during the dry and wet season. Seasonality seemed to be less evident as water
quality was more likely associated with the type of water sources. The study reports that piped
water connections were free of fecal contamination. Interestingly, the fecal coliforms counts in
unprotected riverside wells and springs were lower as compared to protected dug wells during
the dry season. Despite that protected wells in most countries were less contaminated from fecal
coliforms, the authors argued that the possible explanation for this finding might be related to
the structures of the unprotected wells. While the unprotected wells were shallower and served
several families leading to a rapid turnover which would remove the contaminants
23
(decontaminates itself) frequently, the protected wells were emptied only once or twice a year
(Sandiford et al. 1989).
Other water quality assessments based on water sources typology indicated that the quality of
drinking water is highly influenced by water source types. In particular, Haylamicheal & Moges
(2012) studied the physiochemical and microbial quality of the water for 28 randomly selected
community water sources (14 on-spot springs and 14 dug wells fitted with a hand pump) in
Wondogenet district of southern Ethiopia. The study found that water quality met the WHO
drinking water guidelines in terms of pH, temperature, fluoride, chloride, and turbidity but not
the standard for total and fecal coliforms. Of the total sample, 25 percent of water sources were
contaminated with E.coli while more than 85 percent the samples were contaminated with total
coliforms.
In addition to types of water sources, existing studies also emphasized the role of storage
behavior on water quality at the POU (Clasen & Bastable 2003; Crampton & Aid 2005; McGarvey
et al. 2008; Rufener et al. 2010; Baker et al. 2013). Among the earlier studies on water quality,
Clasen and Bastable (2003) report that 92.2 percent of storage drinking water were contaminated
with fecal matters, and using the case of Bamoko, Mali, Baker et al. (2013), the quality of drinking
water was highly affected by household storage behavior although most households had access
to piped tap water, mainly due to lower concentration of free residual chlorine below the
required level during the storage period.
Studies show that water collection container and water handling practices also affect household
water quality. A study that aims to examine the relationship between water handling practice
and microbial water quality in Addis Ababa, Ethiopia, finds that 34 percent of the samples were
contaminated with fecal coliforms out of the 127 total water samples tested (Crampton & Aid
2005). POU water samples were more contaminated with fecal matters (37%, n=54) than water
samples from sources (33%, n=72). The study has also shown that ‘dip’ methods of water storage,
such as bucket and clay vessel, is more prone to frequent contamination but contamination level
is lower as compared with ‘pour’ methods of water storage such as jerrycan and jug. Narrow-
mouthed storage containers are the safest method of water storage but it may be often difficult
to properly clean them after emptying. They usually store bacteria in the ‘biofilm’ and allow
micro-organism to grow on their surface. Crampton and Aid (2005) therefore suggest that “either
a covered bucket with a floating cup used simply to decant water into another glass for
consumption; or a large yet handheld jug with a lid which can be raised for cleaning” could be a
better solution.
Generally, the microbial quality of drinking water substantially deteriorates along the chain from
source to mouth after collection from improved sources (Clasen & Bastable 2003; Rufener et al.
2010; Wright et al. 2004). Clasen and Bastable (2003) examined the level of thermotolerant
coliforms (TTC) for 100 storage drinking water samples and 20 water source points from which
24
the households draw their drinking water in the Kailahun district of Sierra Leone. The authors
find higher TTC loads both at the point of unimproved sources and at household storage.
Moreover, 92.9 percent of water samples from storage were contaminated with fecal matters
although there were no detectable fecal coliforms per 100ml water samples from improved
water sources. Rufener et al. (2010) found similar results in Bolivia. The authors analyzed 347
water samples taken from different water source points, transport vessels, treated water and
drinking water cups from 81 households, and the findings indicate that fecal contamination
(E.coli) of drinking water considerably higher along the chain from the water sources to the
drinking cups. Furthermore, Wright et al. (2004) arrived at the same conclusion after
systematically reviewing studies on microbial contamination of water between source and POU.
In summary, existing empirical studies suggest that, since water quality is often compromised
during household collection, transportation and storage, water quality protection at the POU
should be as highly emphasized as at the point of source (POS).
3.3 Methods and Data
3.3.1 Data
a) Storage water sample collection
In addition to water sample quality-testing from household storage, information about how
households handle and store their drinking water, and any additional water treatment behavior
were recorded. Drinking water samples were collected to analyze the microbial quality of the
water for a random sample of 454 households using a portable water test kit (a product of
Wagtech WTD, UK) in the field. Using a membrane filtration technique, the test kit detects the
presence of the E.coli bacteria which indicates a recent fecal contamination of the water.8
Storage water samples were kept in coded glass bottles which were properly sterilized using
autoclaves in the local health centers at a temperature of 121 degree Celsius for 30 minutes.9
Water sample tests from household’s storage were conducted between February and March
2014—which is between the end of winter and the beginning of spring seasons in Ethiopia and
considered as a dry period.
b) Community water sources samples
In addition to storage water samples, 61 water source points were tested for the presence of
E.coli. These water sources were selected based on the number of households they serve.
Typically, these water source points serve many households of the villagers. Inaccessibility and
8 As stated in the WHO/UNICEF drinking water guidelines, the number of fecal coliform bacteria (E.coli) in drinking
water samples ideally should be zero. Therefore, in a sample volume of 100ml water, a count of zero E.coli CFU is an indicative a microbiologically safe water supply. If the count exceeds 1 E.coli bacteria (CFU/100ml water), contamination is indicated and appropriate action is urgently required.
9 Enumerators asked household members (usually an adult woman) the following question, “could you please give me some water for drinking” so that their behavior would not be altered.
25
resource constraints were the major factors that hindered the uptake of water samples from all
community sources.
c) Microbial quality of drinking water samples
There are various microbiological indicators of drinking water quality. Several studies used total
coliforms, fecal thermotolerant coliforms, and E.coli bacteria to analyze microbial water quality.
In this study, we analyzed the level of E.coli bacteria (CFU/100ml) water sample because testing
for all known pathogens is a complicated and expensive process in the study areas. Besides, E.coli
bacteria are considered as the best microbial quality indicator of drinking water for public health
protection (Edberg et al. 2000).10
Coliform bacteria are grouped into two categories. Total and fecal coliforms based on their
origins and characteristics. Total coliforms are a group of bacteria widely found in the
environment such as in water and soils as well as in human and animal feces; while, fecal
coliforms are found only in animal and human feces. They are often used to detect and estimate
the level of fecal contamination of water and used to indicate the presence of a health risk. For
instance, the presence of fecal coliforms and E.coli in drinking water indicates recent
contamination of water by human or animal feces and may indicate the possible presence of
other diseases causing organisms such as certain bacteria, viruses or parasites. These pathogens
can cause diarrhea, cramps, nausea, and headache and therefore may pose a special health risk
mainly for infants and children.11
Water samples were collected from household storage for all selected households and the
collected water samples immediately were placed into the portable test kit on-site for
incubation.12 The bacteria colonies allowed to culture on MLSB media which are specifically
formulated to facilitate the growth of E.coli bacteria and prevent the growth of other micro-
organisms. As bacteria are extremely small they should be grown on nutrient plates so that they
can multiply rapidly and become visible for enumeration. E.coli concentrations are reported as
colony-forming units per 100ml (CFU/100ml) of water sample.
Immediately after the water samples were collected, the growth pads dispensed into a sterile
petri-dish and a dissolved media solution was poured over the growth pad. Then the water
sample was filtered through the membrane. When all the 100ml water has been filtered, we
placed the membrane on top of the pad which has been saturated with the MLSB media. In the
next stage, we replaced the petri-dish lid and label with sample identification number and time
10 The identification of E.coli bacteria from contaminated water is not complicated and the results are obtained
relatively quickly and cheaply, even though they are only an indicator of fecal contamination. 11 http://water.epa.gov/drink/contaminants/#five (accessed on March 01, 2015). 12 It is recommended that the time between water sample collection and analysis not to exceed 6 hours and it is one
of the strength of this work that we could able to perform the test on the field immediately after collecting the samples from household storage.
26
and placed the petri-dish into the petri-dish rack. Finally, we placed the filled rack into the
incubator and incubate the samples for a maximum of 24 hours at a temperature of 44 degree
Celsius. Upon completion of the incubation period, we enumerated the number of E.coli
CFU/100ml water. In a membrane filtration method, accurate enumeration of bacteria colony is
difficult when the fecal coliform bacteria counts are greater than 200 CFU/100ml water.13
d) Descriptive analysis
The descriptive statistics about the respondent’s background characteristics and socio-
demographic variables are presented in Table 3-3. Out of the 454 surveyed households, 277
households belong to Fogera and 177 households belong to Mecha district. The survey finds that
average household size is about six persons per household which is higher than the reported
mean household size of 5 in rural areas by Central Statistical Agency of Ethiopia (CSA & ICF
International 2012). The survey finds that literacy level (for reading and writing in the local
language) is 9 percent for primary caretakers and 44 percent for household heads. Only a few
have completed primary school indicating that most respondents in this study are illiterate.
Based on the JMP classification; the survey shows that 50 percent of the households get their
drinking water from improved water sources, such as protected wells/springs, about 37 percent
obtain water from unprotected wells/springs and the remaining 13 percent of the households
depend on surface water sources (Table 3-3). The proportion of households having improved
water source is similar to the WHO/UNICEF 2015 progress report; however, the use of surface
water is relatively lower in our sample (WHO/UNICEF 2015) (12% compared to 16%). On the other
hand, the WHO/UNICEF 2015 report indicated that 28 percent of rural Ethiopian households have
access to improved sanitation facility, but our result shows that access to this service is virtually
non-existent in the study areas which is quite surprising. Although 42 percent of the households
reported that they have a simple pit latrine, they may not use it frequently. Many of these latrines
were constructed in response to a push by the local governments. It is a common practice for
most women to go to the bush/open field early in the morning and late in the night for
defecation. The survey also revealed that more than 76 percent of the primary caretakers
defecate without a toilet before the survey. Open defecation is a norm and practiced by most
rural households. More than 57 percent of households in the study areas practice open
defecation, which is much higher than the rural national average open defecation rate of 43
percent (WHO/UNICEF 2014). The study also finds that only 5 percent of the households have
access to protected drinking water source in their own yard/premises, and more than 84 percent
of households on average spend about 25 minutes for a round water collection trip. Moreover,
about 34 percent of the households need round trip of 30 minutes or more to obtain drinking
water from the sources. This suggests that the proportion of households that spend more than
13 For accurate enumeration of bacteria, the number of fecal coliform colonies on the membrane following
incubation should be in the range of 20−200, and more than 200 colonies are difficult to count precisely.
27
30 minutes for a round trip for water collection are lower than what is indicated by the CSA and
ICF International (2012) report (34% compared to 62%).
Table 3-3: Descriptive statistics – household and community characteristics (N=454)
Variables Description Mean SD
Demographic characteristics Household head age Age in years 37.72 8.64 Household head literacy 1=read and write; 0= otherwise 0.44 0.49 Primary caretaker age Age in years 30.33 6.64 Primary caretaker literacy 1=read and write; 0= otherwise 0.09 0.29 Highest education The highest grade completed in a household 3.50 3.05 Number of adult females Female household members aged >14 years 1.22 0.49 Household size Number of household members 5.98 1.77 Household density Number of people living per room 3.30 1.27 Housing and household possession Roofing materials
Corrugated iron sheet 0.91 0.28 Thatch 0.09 0.28
Assets value Total asset value excluding livestock in 1000 Birrb 5.88 6.05 Water, sanitation and hygiene Primary drinking water sources
Private-protected dug wells 0.05 0.22 Shared-protected dug wells/spring 0.44 0.50 Unprotected dug wells/spring 0.39 0.49 Surface water 0.12 0.32
Minutes to water sources a Time needed for a round trip water collection 24.18 14.19 Stored drinking water quality 1=Contaminated with 1 or more E.coli 0.58 0.49 Household water treatment Treating household drinking water (1=yes) 0.08 0.27 Water collection container 1=Jerrycan; 0=Clay vessel 0.83 0.37 Handwashing with soap Handwashing with soap by primary caretaker during
handwashing demonstration (1=yes) 0.27 0.45
Pit latrine Households with a pit latrine (1=yes) 0.42 0.49 Garbage disposal
Burning/Buried 0.11 0.31 Throw-away in the yard 0.54 0.50 Throw-away outside the yard 0.13 0.34 Used as a fertilizer 0.22 0.42
Agriculture Irrigation Practicing irrigation farming (1=yes) 0.66 0.47 Livestock holding Total livestock holding in Tropical Livestock Units 3.97 1.87 Community characteristics Water user association (WUA) Presence of WUA in a village (1=yes) 0.29 0.46 Distance to health center Distance to the nearest health center in km 4.97 4.09 a The mean is calculated for households whose water sources are not in their own yard/premise. b The exchange rate during the time of the survey was 1 Euro = 26.02 Ethiopian Birr.
Source: Author’s compilation using survey data.
28
3.3.2 Empirical estimations
To examine the determinants of microbial quality of storage drinking water, socio-demographics,
water sources, as well as collection, storage, sanitary and waste disposal behaviors were assessed
using simple chi-square analysis followed by a multivariate regression analysis. Admittedly, due
to the collinearity among the variables and the cross-sectional nature of the data, our analysis is
constrained to make any causal interpretation of the results. We instead investigate the degree
of correlation between the microbial quality of storage water and socio-demographic, water
sources, and sanitary factors.
In the multivariate analysis, we examined two different measurement specifications for the
dependent variable, that is, water quality. First, the dependent variable indicates the number of
E.coli (CFU/100ml water). We transformed the dependent variable (E.coli counts) into the inverse
hyperbolic sine (IHS), which is defined as: 𝑖ℎ𝑠(𝑦) = log(𝑦 + 𝑠𝑞𝑟𝑡(𝑦2 + 1) where Y is the number
of E.coli and estimated using OLS.14This transformation is an alternative to log transformation
when the dependent variable takes zero values (MacKinnon & Magee 1990) and we interpret the
coefficients of the explanatory variables same as the log transformation. Second, we measured
the dependent variable as a binary outcome, which indicates the presence or absence of E.coli
CFU/100ml water, that is, 𝑦 is equal to 0 if E.coli is less than 1 and 𝑦 is equal to 1 if E.coli is greater
than or equal to 1, and estimated using maximum likelihood estimator in the subsequent
analysis.
3.4 Results and Discussion
3.4.1 Bivariate analysis
This section highlights the relation between socio-demographics, water sources and water
handling, or household characteristics and quality of storage drinking water.15 More than 58
percent of stored drinking water samples were contaminated with E.coli.16 This result is not
surprising when compared to earlier findings elsewhere in Ethiopia. For instance, a study in Kersa
district of Eastern Ethiopia found that more than 78 percent of sampled households’ storage
water were contaminated with E.coli (Mengistie et al. 2013; Tsega et al. 2014). In the bivariate
14 The reason for this transformation is that we cannot take the normal log of y as we have many observations with
zero value, and the distribution of E.coli is positively skewed because coliforms naturally grow exponentially. 15 The influence of socio-demographic variables, such as primary caretaker’s and household head education and
household size, turns out to be insignificant. The non-significant bivariate association between water quality and education variable might be partly explained by the low level of actual school attainment among surveyed households (table not presented here).
16 The presence of E.coli colony units on storage drinking water of the surveyed households ranged from 0 to 195 (CFU/100ml) water sample.
29
analysis, water quality indicator is measured as a dummy variable (the variable is equal to 1 if 1
or more E.coli CFU/100ml water sample is counted, otherwise 0).
Water sources, handling and collection, and storage water quality
The relationships between water sources, collection and handling practices and storage water
quality have been presented in Table 3-4. The results show that types of water sources, types of
water collection containers and garbage or waste disposal patterns have statistically significant
influence on storage water quality.
We found that types of water sources are highly associated with storage water quality.
Households who had so called ‘improved’ water sources showed much better microbial water
quality than households who had either unprotected dug wells/springs or surface water sources.
Kremer et al. (2009) shown that simple spring protection significantly improves the microbial
quality of both POS as well as POU water. The result in Table 3-4 also shows a significant
association between the types of water collection containers and storage water quality
(p=0.000). More than 83 percent the households use jerrycan for water collection. Most
households use the same jerrycan for both hauling and storing drinking water at home. However,
household water storage and treatment practice do not have a significant influence on storage
water quality. Although half of surveyed households get their drinking water from unimproved
source, the proportion of households applying any form of water treatment is low (8%). This
clearly indicates a lack of awareness of the need to treat household drinking water among rural
households in the country. About 24 percent of households also reported that they do have a
separate water storage containers.
Moreover, the proportion of households with contaminated water with E.coli was lower among
households who had simple pit latrine than those who did not have and the relationship is
significant. Similarly, households in which the primary caretaker washes her hands with soap had
better storage water quality than households whose primary caretaker did not. Hands may come
into contact with feces as a result of multiple factors and pose a potential risk of contaminating
household water during water handling (Trevett et al. 2005). Safe disposal of household’s
garbage greatly influenced household water quality (p<0.000). Solid garbage disposal methods in
the study areas include dug-out/burning (11%), throwing away in the yard (54%), throwing-away
outside the yard (13%), and composting or used as a fertilizer (22%). Although a higher
percentage of non-irrigator households had better water quality than irrigator households, the
relationship is not statistically significant.
30
Table 3-4: Relationships between water sources, collection and stored drinking water quality
Variables
N Water quality (%) Chi-squared
(χ2) P-values
Contaminated Uncontaminated
Water sources Private protected dug wells 23 43.48 56.52 41.640 0.000 Shared protected dug well/spring 202 43.07 56.93 Unprotected dug wells/spring 176 72.16 27.84 Surface water 53 75.47 24.54 Water collection containers Jerrycan 379 62.01 37.99 14.014 0.000 Clay-vessel 75 38.01 61.33 Household water treatment Yes 35 71.43 28.57 2.748 0.097 No 419 57.04 42.96 Water storage Yes 111 61.26 38.74 0.585 0.445 No 343 57.14 42.86 Handwashing with soap Yes 124 47.58 52.42 7.831 0.005 No 330 62.12 37.88 Household sanitation facility Pit latrine 189 51.85 48.15 5.277 0.022 No facility (open field/bush) 265 62.54 37.36 Garbage disposal Burning 49 16.33 83.67 59.309 0.000 Throw-away in the yard 245 71.43 28.57 Through away outside the yard 59 42.37 57.63 Used as a fertilizer 101 55.45 44.55 Irrigated agriculture practice Yes 302 58.94 41.06 0.232 0.630 No 152 56.58 43.42
Source: Author’s compilation using survey data.
Community water sources quality
Of the total 61 community water source samples tested, 73.8 percent of the total samples were
contaminated with E.coli. Of the water samples collected, 58.6 percent of protected dug
wells/springs, 84.6 percent of unprotected wells/springs and 100 percent of surface water
sources were contaminated with E.coli. Forty-eight percent of the samples were from protected
wells/spring while the remaining were from unprotected wells/springs and surface water (Table
3-5). Protected wells/spring has lower E.coli concentration (CFU/100ml water) than unprotected
wells/springs and surface water sources. The finding is evident that most community water
sources are of unacceptable microbial quality for household consumption unless water is made
safer. The presence of rampant drinking water contamination both at the POS and POU can pose
a high risk of public health problems from water-related diseases.
31
Table 3-5: Community water source sample test results
Contaminated water sources
Source type N Column percentage Row percentage Mean E.coli per 100ml
Protected wells/spring 29 37.78 58.62 6.83
Unprotected wells/spring 26 48.89 84.62 34.46
Surface water 6 13.33 100 61.33
Total sample 61
Source: Author’s estimates using survey data.
As Figure 3-2 illustrates, about half of the surveyed households reported that they do not have
any major problems related to their primary drinking water source. However, one-quarter of the
households stated that the water they collect from their primary drinking water source is of poor
quality. Traveling long distance and spending considerable time for water collection and irregular
water supply also reported as major problems in the study areas. Moreover, due to water
unavailability, some water points are only open during certain times of a day and households
may be forced to collect water from other unimproved sources.
Figure 3-2: Major problems related to drinking water supply in the study areas
Source: Author’s computation using survey data.
In many developing countries, particularly in the sub-Saharan region, women and girls bear the
burden of water fetching for household uses and often needs to travel more than half an hour
round trip (WHO/UNICEF 2010b). Figure 3-3 shows which household members usually collect
216
112
8263
0
50
100
150
200
250
No problem Water is dirty Source supply isirregular
Source is too far
Nu
mb
er o
f h
ou
seh
old
s
32
water for households whose primary drinking water source is not on premises/own yard. Often
females are more responsible for household water collection than other household members in
the study areas. For instance, adult women are approximately 10 times more likely to collect
water for household consumption than adult men. This result is consistent with a recent finding
from the Demographic and Health Survey (DHS) report (CSA & ICF International 2012). As the
time burden of domestic water collection is primarily borne by adult women and school-age
female children, it has other implications—such as gender equality, social empowerment, and
school attendance—especially for girls. The situation is similar in many other developing
countries (Sorenson et al. 2011). Therefore, the provision of clean and adequate WATSAN
services foremost benefits women and children—because it reduces the burden of traveling long
distances to fetch water, which in turn increases their time to participate in community activities
and on girls to go to school.
Figure 3-3: Household members who usually collect when water source is not on premises
Source: Author’s computation using survey data.
3.4.2 Multivariate analysis
This section discusses the empirical results from the multivariate regression. The OLS regression
results are presented in Table 3-6 while the logistic estimated odds ratios are presented in Table
3-7. The OLS model was used to determine the factors associated with the natural logarithm of
E.coli CFU/100ml water, that is, the intensity of fecal contamination. Given that drinking water is
generally of poor quality among the sampled households, this approach allows us to investigate
the incremental effects of the covariates on the level of E.coli concentration. On the other hand,
the logistic regression was used to estimate the odds of contaminated water, that is, the binary
outcome of potable or unpotable water (1 if 1 or more E.coli CFU/100ml water is counted,
otherwise 0). For both types of regression analysis, we estimated different model specifications
in stages to control for the potential confounding factors.
356
138
56 360
100
200
300
400
Adult women School-agefemale children
School-age malechildren
Adult men
No
. of
ho
use
ho
lds
33
The OLS regression results presented in Table 3-6 show that types of primary water sources
influence storage water quality. Household storage water from protected wells/spring had lower
E.coli (CFU/100ml water) compared to unprotected wells/springs and surface water sources—
implying that water from unprotected wells/spring and surface water sources had significantly
higher level of E.coli than protected sources (model 2, Table 3-6). This association remains
significant after further adjustment for household demographic characteristics. However, the
pattern of relationship between water sources and E.coli level of storage water does not remain
the same after controlling for sanitary characteristics. The result suggests that water from
unprotected wells/spring had higher level of E.coli than other alternative water sources, and the
difference between water from protected and surface water sources is not statistically significant
(model 3, Table 3-6). Similarly, the results from the logistic regression estimates presented in
Table 3-7 suggest that water from surface water is 3.7 times more likely to be contaminated with
fecal materials compared to protected wells/springs; however, this odds disappear after
controlling for sanitary factors (model 3, Table 3-7). On the other hand, water from unprotected
sources is 2—3.6 times more likely to be contaminated than protected sources.
The time to walk to a water source is highly positively associated with the level of E.coli. Traveling
long distance to collect water increases the risks of the water to be contaminated. This
relationship remains strong after controlling for household demographic and sanitary
characteristics. Available evidence indicated that water collected from improved sources may be
re-contaminated during collection, transportation, and storage (Wright et al. 2004). The
association between household water collection container vessels and the level of E.coli is strong
even after adjusting for household’s socio-demographic and sanitary characteristics. Households
who use jerrycan container for water collection activities had higher E.coli level than households
using clay vessel. In rural Ethiopia, it is found that more than 95 percent of households do not
adequately and regularly clean their water jerrycan containers (Kinfegabriel 2014).
Regarding household demographic characteristics, the household level of education, as
expected, negatively affects the level of E.coli in storage water. Furthermore, household density
is strongly positively associated with storage water level of E.coli across all model specifications.
On the other hand, the number of adult female household members positively influence storage
water quality but the association vanishes after controlling for household sanitary conditions.
Our study also indicates that household’s methods of garbage disposal patterns are highly
associated with storage water level of E.coli.
In the third model specification, although the relationship between latrine and level of E.coli is
not statistically significant, it is positively associated with the level of E.coli when we introduce
the interaction terms between latrine and water source location (Table 3-6). This implies that
availability of pit latrine may increase the risk of fecal contamination of storage water if a water
source is located on premises. This seems intuitive if ground water sources are inadequately
34
protected and/or located close to a pit latrine.17 Our results also suggest that handwashing with
soap and household assets (a proxy for wealth) are negatively associated with storage water level
of E.coli, and similar results are shown in the logistic model specification.
Most households practice mixed farming and often livestock lives together with human beings.
That is, more livestock unit means more crowded living conditions in rural areas. The negative
relationship is expected and the effect size is relatively large. Households engaged in irrigated
agriculture have low storage water quality. As irrigated agriculture has complex interactions with
drinking water, household water can easily become affected through agricultural practices or
through multiple water use. The existence of water user association (WUA) in the
community/village is also robustly associated with better storage water quality. Finally, the r-
squared for the OLS regression is modest for a cross-sectional study and it ranges from 0.17 to
0.46 when we adjusted for socio-demographic and sanitary characteristics.
In most cases, the two regression tables produce similar results with expected signs. For instance,
across all model specifications, time to water sources and household level of education are
significantly associated with lower odds of fecal contamination while water collection jerrycan,
household density, and garbage disposal patterns are significantly associated with higher odds of
fecal contamination (Table 3-7). Moreover, households who keep livestock is associated with 54
percentage points increase in the odds of contaminating their storage water. However, contrary
to the OLS regression results presented in Table 3-6, some of the variables, such as irrigation
practice, water source location and its interaction effect with pit latrine, which greatly influence
storage water level of E.coli at different levels, do not have a statistically significant influence in
the logistic regression results (Table 3-7). This indicates that these variables could be proximate
causes for poor storage water quality as their effects depend on the distribution of E.coli
(CFU/100ml).
17 As the interaction term is not statistically significant in the logistic regression, we did not control for it.
35
Table 3-6: Estimates from OLS regression predicting the natural log of E.coli
VARIABLES Model 1 SE Model 2 SE Model 3 SE
Primary drinking water source a Unprotected well/spring 1.040*** 0.186 1.052*** 0.164 0.335** 0.152
Surface water 1.190*** 0.261 1.127*** 0.238 0.212 0.241
Minutes to water source (1=30min/ less) -1.061*** 0.242 -0.981*** 0.225 -0.868*** 0.224
Container (1=Jerrycan) a 1.197*** 0.198 1.146*** 0.191 1.048*** 0.177
Highest education comp. -0.081** 0.034 -0.034 0.026
Household size -0.002 0.068 -0.081 0.057
Household density 0.398*** 0.082 0.310*** 0.067
Number of adult females -0.408** 0.178 -0.187 0.135
Garbage disposal methods a
Throw in the yard 1.480*** 0.230
Throw away outside the yard 0.559* 0.310
Used as fertilizer 0.860*** 0.245
Handwashing with soap (dummy) -0.547*** 0.161
Log of assets value -0.335*** 0.099
Livestock holding 0.275*** 0.059
Irrigating households (dummy) 0.532*** 0.134
Water user association (dummy) -1.374*** 0.180
Pit latrine (dummy) -0.025 0.172
Water source location (1=on premises) -0.496** 0.236
Water source location X latrine 1.070*** 0.364
Constant 1.520*** 0.281 0.970** 0.389 2.714*** 0.816
Observations 454 454 454
R-squared 0.17 0.25 0.46
Model F-stat 35.28 34.17 61.32
Model P-value 0.000 0.000 0.000 a Omitted reference categories are protected well/spring, clay vessel, and dugout/burning. Robust standard errors adjusted for clustering at the village level are in parenthesis; Significance *** p<0.01, ** p<0.05, * p<0.1
Our results generally indicate a common problem of poor storage water quality in the rural areas
of Fogera and Mecha district with more than 58 percent of the households having at least 1 E.coli
CFU per 100ml water. Our results further suggest that storage water quality is strongly associated
with water sources, water collection time and types of containers, the existence of WUA in a
village, household demographic structures, and households overall sanitary characteristics. Our
findings are also consistent with other studies that reveal substantial levels of fecal
contamination of storage water after collection from improved sources that are less prone to
high level of fecal contamination (Wright et al. 2004). Existing underdeveloped rural water
infrastructure couples with poor household water quality and lack of key hygiene practices pose
a substantial risk of waterborne infectious diseases in the country.
36
Table 3-7: Estimates from Logistic regression predicting E.coli contamination (1 if E.coli >= 1)
VARIABLES Model 1 SE Model 2 SE Model 3 SE
Primary drinking water source a
Unprotected well/spring 3.246*** 0.717 3.582*** 0.752 1.958** 0.550
Surface water 3.693*** 1.277 3.619*** 1.213 1.161 0.433
Minutes to water source (1=30min/less) 0.387*** 0.116 0.396*** 0.123 0.373** 0.160
Container (1=jerrycan) a 2.635*** 0.759 2.702*** 0.769 3.470*** 1.236
Highest education completed 0.901** 0.037 0.909** 0.038
Household size 1.031 0.080 0.904 0.085
Household density 1.428*** 0.165 1.466*** 0.181
Number of adult female 0.655* 0.160 0.758 0.213
Garbage disposal methods a
Throw in the yard 14.755*** 7.629
Throw away outside the yard 2.869* 1.765
Used as fertilizer 5.770*** 3.245
Handwashing with soap (dummy) 0.394*** 0.120
Log of assets value 0.734** 0.101
Livestock holding 1.571*** 0.176
Irrigating households (dummy) 1.640* 0.435
Water user association (dummy) 0.157*** 0.055
Pit latrine (dummy) 0.981 0.250
Constant 0.795 0.277 0.471 0.243 0.672 0.910
Observations 454 454 454
Pseudo R-squared 0.10 0.15 0.35 Model Chi2 55.79 105.46 170.34 Model p-value 0.000 0.000 0.000
a Omitted reference categories are protected well/spring, clay vessel, and dugout/burning. Robust standard errors adjusted for clustering at the village level are in parenthesis; Significance *** p<0.01, ** p<0.05, * p<0.1; and coefficients are odds ratio (OR).
In general, rural water supply infrastructure often does not guarantee the basic safe water supply
indicators defined by the WHO such as water quality, quantity, continuity, and coverage or
accessibility. It is estimated that, at any given time, one-third of rural water supply schemes are
non-functional (MoWE 2007). A recent survey of 57 diverse water source schemes also showed
that 38.6 percent of the systems were non-functional on the day of the visit (Welle & Williams
2014). For instance, a community has to wait may be for more than a month to get broken hand
pumps repaired. As a result of the intermittent and unreliable water supply, most households are
forced to collect water from other unimproved water sources as people generally prefer the taste
of spring water than constructed wells in the study areas (Beyene 2012). Besides, water from
spring sources requires less waiting time than water from constructed wells.
In the bivariate analysis the influence of household water treatment practice on storage water
quality is not strong (Table 3-4). As we have discussed earlier, the weak relationship between
37
household water treatment practice and storage water quality is because of lack of regular use
of any form of water treatment in our sample households. For instance, among the households
who use some form of water treatment, more than 80 percent of these households applying
chlorine-based methods, of which 72 percent households use this method during the month
before the survey. None of the respondents reported regularly treating their household drinking
water. However, we observed that E.coli levels are significantly lower for households practicing
any water treatment methods compared to households who did not. The empirical evidence that
household water treatment and safe storage (HWTS) practice in improving the microbiological
quality of drinking water is well-documented (for example; see Clasen 2015; Fewtrell et al. 2005;
Mengistie et al. 2013; Mintz et al. 1995). Largely, the habit of treating water before drinking is
critically slim in both urban and rural households. This is because of a lack of awareness about
domestic water quality and its health consequences, people often perceive that clean water is
‘clean’ as long as it is not turbid. For instance, about 87 percent of urban households and 91
percent of rural households do not practice any form of household water treatment in the
country (CSA & ICF International 2012).
The types of household storage container can also influence household water quality (Levy et al.
2008; Günther & Schipper 2013). In this study, types of water collection containers are
significantly associated with the quality of water consumed by the household. More than 83
percent of the households identified jerrycan as a favorite container for hauling and storing their
drinking water and 24 percent of households had separate water storage containers. Our result
shows that jerrycan increases the risk of storage water contamination, and this is mainly due to
inadequate cleaning. Although jerrycan container has an advantage of being narrow-mouthed,
rural households do not properly clean it. Cleaning its inside part with a simple washing is
challenging. A study in rural Ethiopia reported that more than 95 percent of households do not
adequately and regularly clean their water container or jerrycan (Kinfegabriel 2014). Previous
studies elsewhere showing that storage container characteristics, such as narrow versus wide
mouth and covered versus uncovered, are key factors in determining storage water quality
(Mintz et al. 1995). It is assumed that water pouring is safer than dipping but this research
questioned if a narrow-necked container, such as jerrycan, is the safest methods of water
storage. Households opt to store water for future use when the water supply is unreliable and
intermittent. However, Brick et al. (2004) suggest that drinking water contamination will also be
higher if water is stored for longer period.
Our results also highlighted that increased water collection time increases the risk of storage
water contamination. This is in line with studies showing that the microbiological quality of water
obtained from improved sources significantly deteriorates during collection and transportation
(Wright et al. 2004). Moreover, water collection time determines the quantity of water a given
household can collect and consume (Cairncross 1987), which is a critical determinant of key
38
hygiene practices (Cairncross 1997; Curtis et al. 2000; Gilman et al. 1993). On the other hand,
more time allocation for household water collection may allow households to collect sufficient
water and to maintain key hygiene practices, such as washing hands at critical times, which can
influence storage water quality (Curtis et al. 2000). For instance, hands may become into contact
with feces through various mechanisms and become contaminated with water. This implies that
proper hygiene practices can reduce storage water contamination.
Household demographic variables, particularly household density are strong predictors of
storage water quality. It can be argued that crowded living conditions might influence the overall
hygiene and sanitation environment that probably increase the risk of storage water
contamination. It is also a common understanding that the level of E.coli in storage water is
expected to positively correlate to household size due to possible contacts from the many hands
to water containers, but the effect of this variable turns out to be statistically insignificant.
Moreover, higher household education is expected to correlate with a better understanding of
water quality and sanitary behavior, which in turn could influence household water quality
through improved water handling and hygiene practices. However, our results show that the
effect of education is small, or even statistically insignificant (model 3, Table 3-6). This could be
explained by low levels of school attainment, or suggest that the primary caretaker’s level of
education might be more important in determining stored water quality than any other
household members.
On the other hand, pit latrine availability increases the level of E.coli on storage water for
households who use well water sources in their own premises. Megha et al. (2015) showed that
the microbiological quality of ground water deteriorates where pit latrines are placed close to
the source. In support of this, our result shows that households having a pit latrine and using own
wells located in premises have high levels of E.coli on storage water. In addition to the type of
well, the risk of water quality problem with groundwater supplies is directly related to how close
it is to potential sources of contamination, that is, the risk of contamination decreases as the
distance between the well and potential contamination sources increases. Therefore, source
water contamination from own latrine could be one possible channel for high contamination of
storage water. Moreover, as private-wells are often shallow and inadequately protected
compared to community hand dug wells, this might increase risks of contamination from
household waste water, animal droppings, flood-washed wastes, dirty well surroundings and
water-drawing buckets.
In rural areas, agriculture and livestock rearing, which are the primary sources of livelihood, have
complex interactions with household water quality. Most households keep livestock and often
livestock lives together with human beings, increasing the risk of household water
contamination. The negative relationship is expected and the effect size is also relatively large.
Households engaged in irrigated agriculture have lower water quality. Where access to improved
39
drinking water is limited, households opt to use irrigation water for domestic purpose, which is
often of poor quality. A significant portion of households reported that they directly withdraw
water from irrigation sources for household consumption. Although irrigation water increases
water availability for domestic purpose, it might increase the risk of storage water contamination.
Similarly, household members working on an irrigation field may come into contact with
domestic water and contaminate if proper personal hygiene and handwashing are lacking.
We also expect households with more accumulated assets (wealth) to live in a more sanitary
environment, which can reduce drinking water contamination. Livestock ownership may offset
the net positive gains of household assets on stored water quality, as livestock can be a source
of pathogens which directly affect the water quality. Moreover, household assets are moderately
correlated with livestock ownership, irrigation practices, and household education outcomes.
This suggests that most socioeconomic variables can influence water quality indirectly through
various pathways.
Another interesting finding is that the existence of WUA in the village influences the quality of
storage water quality. Households that belong to a community in which there is a WUA reported
better storage water quality. This association is consistently statistically significant across all
model specifications. The WUA is primarily responsible for monitoring, supervising and handling
conflicts among household users of community water sources. The influence of WUA on storage
water quality might be via improving the protection of water sources from external
contamination. Basically, WUA were instituted in many villages for governing rural water supplies
when a new water source point was constructed (Tilahun et al. 2013).
Increasing the provision of rural water supply is the agenda of both the regional government and
other development partners, yet, most rural households had to travel long distance that may not
even guarantee them to get improved water sources. Moreover, widespread household water
contamination undermines the progress that has been made in terms of increasing access to
improved water supply in rural areas. Today, lack of access to clean and adequate drinking water
and poor sanitary environment is a critical public health problem in Ethiopia, contributing to 70
percent of the diarrheal diseases burden in the country (FMoH 2005). Unsafe water is not just
dirty; it can be deadly if people drink it without any prior treatment. Therefore, any intervention
that aims at increasing access to safe and clean rural water supply should be accompanied by
large-scale household interventions such as safe water storage, POU water treatment.
3.5 Conclusions and Recommendations
Most of the health problems of children in Ethiopia are communicable diseases due to unsafe
drinking water sources, improper water handling practices, and poor sanitation facilities. The
findings that 74 percent of water sources and 58 percent of stored household drinking water
samples were positive for E.coli bacteria indicated that majority of the rural population is at high
40
risk of waterborne diseases. About 50 percent the surveyed households get their water from
protected or ‘improved’ water sources; however, more than 42 percent of these households’
storage drinking water was contaminated with fecal materials. The findings indicated the
rampant drinking water problems both at POS and POU in rural areas of Ethiopia. The situation
is almost similar in many other rural areas of the country (Mengistie et al. 2013; Tsega et al.
2014). Moreover, Wright et al. (2004) showed that microbiological quality of drinking water
significantly declines after collection from the acceptable quality of water sources. It is widely
understood that POU water treatment and safe water storage are more effective ways, and
should be a focus of intervention to ensure the quality of water being consumed (Clasen &
Bastable 2003; Gundry et al. 2004).
The study suggests a need to promote water safety along the POS to POU to advance the
Sustainable Development Goal (SDG-6) of ensuring access to clean water for everyone.
Therefore, some interventions can be implemented to address the problem of poor water quality
until the long-term goal of providing clean and safe water supplies for all Ethiopian rural
households can be achieved. In addition to expanding the WATSAN infrastructures to increase
access for the unserved population, the following recommendations are made to improve the
overall situation of poor water quality both at the POS and POU. First, available water source
points should be adequately protected. Most community water sources considered being
‘improved’ and widely considered to provide safe water showed the presence of E.coli which is
not in compliance with both the national and the WHO guideline standards. Second, promoting
household water treatment methods and products to make water safer. A simple water
treatment practices of drinking water avoids the risk of contaminated water. Household water
treatment and safe storage, such as boiling, filtering, or chlorinating water at home, are effective
in improving the quality of storage water (Clasen 2015). Therefore, promoting HWTS and the
health risks of drinking contaminated water may bring significant progress in the provision of
clean and safe water. Third, as most of the households use jerrycan for water collection and
storage, either providing safer and convenient storage containers or promoting how to clean it
properly would avoid substantial risk of water contamination. Moreover, since adult women and
school-age children are disproportionately responsible for household water collection, targeting
these age groups in any hygiene education intervention on handwashing, water collection and
storage may generate substantial improvement in storage water quality. Fourth, ad hoc water
quality-testing and quality control mechanisms for rural water supply systems need to be in place
to ensure safety of drinking water supplies. Determining the public health risk associated with
drinking water quality is useful; however, in practice monitoring of pathogens is generally not
carried out either systematically or regularly. Once the water supply infrastructure is in place,
systematically well planned and designed sanitary management needed to ensure safe drinking
water. Fifth, Private-well water sources should not be developed close to household’s latrine to
prevent seepage. The quality of well water often is directly related to the care taken in well
41
construction. Many of the private-water sources in the study areas are bucket wells and they are
often shallow and under-protected and can be easily contaminated by latrine, animal droppings,
dirty ropes and buckets and households waste. Therefore, one should plan carefully before
choosing the site to minimize the risk exposure from external contamination as its location
determines the quality of the water obtained. Six, building the capacity of WUA and providing
training in water source protection, environmental sanitation, and systems operation and
maintenance. The provision of drinking water in rural areas through community water scheme is
the conventional way and this is the only existing alternative to increasing access to clean water.
Supporting WUA to enable them to repair and manage available water sources is, therefore,
critical in the provision of sustainable rural water supply.
On top of that, variations in the community and household behavioral and sanitary factors are
key determinants of household storage water quality. Unsafe sanitation habits, inadequate
garbage disposal, and widespread open defecation could be the primary causes of drinking water
contamination. Furthermore, keeping livestock units separately from household dwellings and
out of the water source catchment areas can improve water quality. Therefore, without proper
waste disposal and sanitation facilities, water source points are highly prone to gross
contamination from human and animal feces which are the principal sources of disease-causing
pathogens. Generally, the association between improved water supply and safe household water
seems too simplistic, and a mix of instruments needed to address the complex problem of
drinking water safety and to make progressive improvements in the next decade in terms of other
aspects the SDG-6 indicator as well.
42
4. The Impact of Drinking Water Quality and Sanitation on Child Health
4.1 Introduction
Globally, more than 700 million people live without an improved water source, and eight out of
10 of these people live in rural areas (WHO/UNICEF 2014). An estimated 2.5 billion people also
lack access to improved sanitation facilities, of which a billion people practice open defecation
(WHO/UNICEF 2014). Sub-Saharan Africa and Southeastern Asia are the regions with the lowest
coverage of improved sanitation in the world. In parallel, millions of people are suffering
worldwide from WASH related diseases such as diarrhea, skin diseases and trachoma. Inadequate
WASH services are linked to 88 percent of diarrhea cases worldwide and result in more than 1.5
million children deaths each year—mostly under-five children (WHO 2002; WHO/UNICEF 2009).
Ethiopia is the second-most populous country in Africa and about 85 percent of the country’s 96
million people live in rural areas.18 Mortality and malnutrition are serious health problems among
under-five children in the country. Although the country has reduced under-five mortality rate
by more than half since 1990 (UNICEF 2012), Ethiopia was ranked 27th in the global under-five
child mortality rate estimates in 2007 (UNICEF 2009). As per the World Bank (2005) Country
Status Report on Health and Poverty, child mortality in Ethiopia is among the highest in the world;
almost one out of 10 babies born in the country (97 per 1000) do not survive to celebrate their
first birthday, and one in every six children die before their fifth birthday. More than 90 percent
of these deaths were due to preventable diseases such as pneumonia, diarrhea, malaria, measles,
malnutrition and HIV/AIDS (FMoH 2010). As a result of inadequate access to improved WASH
services in both rural as well as urban areas, the spread of diseases caused by them is a major
health problem in the country. It is estimated that 60 percent of the country’s disease burden
and 70 percent of the diarrheal diseases were mainly attributed to poor WASH services (FMoH
2005). It is commonly understood that most of these deadly diseases caused by poor WASH are
largely preventable.
In the last decade, the GoE has made significant progress in increasing access to WATSAN
services, and in modifying long-held hygiene habits. Nonetheless, as shown in Figure 4-1, access
to proper sanitation facility is inadequate—particularly in rural areas. The difference between
urban and rural households is also striking: based on the WHO definition, only 28 percent of rural
households have access to improved sanitation while the remaining households rely on unsafe
sanitation services and open defecation. This issue is compounded by water source
contamination by animal and human feces, which is caused by inadequate protection to most
water sources in rural areas. WHO/UNICEF (2015) estimated that 57 percent of Ethiopian
households have access to improved drinking water sources—93 percent in urban areas and 49
18 http://worldpopulationreview.com/countries/ethiopia-population/
43
percent in rural areas. As a result, most of the rural population rely on unimproved water sources,
such as rivers, lakes, ponds, streams, rainwater, unprotected springs and wells, irrigation water
from canals and dams, as a source of water for drinking and other domestic uses. Moreover, less
than 20 percent of the population is regularly washing their hands with soap and water at critical
times (FMoH 2011). The same report indicated that 8 percent of the population follow the safe
drinking water chain from source to mouth (FMoH 2011). Due to unsafe handling and storage of
drinking water, 40 percent of the domestic water consumption is contaminated with fecal
matters.
The impacts of unsafe drinking water and sanitation behavior on child health outcomes have not
been investigated sufficiently, and little is known about the impacts in the context of rural
Ethiopia. Existing empirical works on the health impacts of WATSAN in Ethiopia did not give due
attention to water quality at the household level and child stool disposal habits (Cameron 2009;
Kirchberger 2008).
Figure 4-1: Percentage of population by type of sanitation facility in Ethiopia, 2015
Source: Author’s compilation using the WHO/UNICEF JMP dataset.
Using a variety of impact evaluation methods, this chapter aims to shed more light on the health
impacts of household drinking water quality and sanitation behavior on under-five children in
two rural areas in Ethiopia—Fogera and Mecha districts. The innovative aspects of this chapter
lie in two key contributions: first, household’s drinking water quality was determined by actually
testing the microbiological quality of water samples taken from drinking water stored in the
27 28
28
40
8
14
2730 29
6
34
29
0
10
20
30
40
Urban Rural Total
Per
cen
tage
of
po
pu
lati
on
Improved facilities Shared facililtes Other unimproved Open defecation
44
household19 using a membrane filtration method rather than looking at the types of household
drinking water sources. The second novelty of this chapter is that we take into account
household’s behavior on child stool disposal, which is critical in child health outcomes. The key
findings are that improved storage water quality and safe child stool disposal highly influence
childhood diarrhea. However, contrary to widely held perceptions, the results suggest that we
should think differently about the health impacts of simple pit latrines in rural areas.
4.2 Water Supply, Sanitation, Hygiene and Health
Improving WASH services has been recognized as a fundamental factor to improve health and as
one of the driving forces of social and economic progress in developing countries. Improving
water quality and sanitation, together with better hygiene practices, can have significant effects
on the health of a population. It reduces the incidence of a variety of waterborne diseases, such
as diarrhea, intestinal helminths, guinea worm, skin diseases, and trachoma (Esrey et al. 1991),
by interrupting or decreasing the transmission of pathogenic disease agents. These health
improvements can, in turn, lead to improved nutritional status and reduced morbidity and
mortality, particularly among under-five children. Table 4-1: presents potential transmission
routes of pathogens and a broader classification of the disease burden associated with unsafe
and inadequate water supply.
Table 4-1: Transmission Routes of Water-Related Diseases
Classification Transmission route Examples of diseases transmitted
Waterborne through ingestion of pathogens in drinking water Diarrheal diseases Enteric fevers, such as typhoid Hepatitis A
Water-washed through incidental ingestion of pathogens in the course of other activities; results from having insufficient water for bathing and hygiene
Diarrheal diseases Trachoma Scabies
Water-based through an aquatic invertebrate host; results from repeated physical contact with contaminated water
Guinea worm Schistosomiasis
Water-related insect vector
through an insect vector that breeds in or near water
Malaria (parasite) and Yellow fever (virus)
Source: Bradley (1977).
Diarrhea is both a waterborne as well as a water-washed disease, and it can be caused by
ingesting water contaminated with human and animal feces which contain pathogenic agents or
ingesting these pathogens directly through various fecal-oral pathways. The latter is likely to
occur when water availability is limited, which hinders proper hygiene practices (e.g., washing
19 Hereafter referred to as storage water
45
hands after defecation). Although diarrhea and malaria are the most prevalent diseases in the
study areas, this chapter focuses on diarrhea in under-five children.
Figure 4-2 shows the fecal-oral routes of disease transmission and how intervention can break
the chain of contamination at various stages of the transmission pathways. Human and animal
excreta are the primary sources of most disease-causing pathogens. As the figure illustrates,
these pathogens are passed from an infected host to a new one via various transmission routes.
They are transmitted via the fecal-oral routes through fluids, hand contact, flies, and food. The
figure also shows the importance of sanitation and safe removal of human feces as a primary
barrier to prevent these pathogens from reaching the domestic environment. Good hygiene
practices and household water treatment also serve as a secondary barrier to prevent the
transmission of disease-causing pathogens. For example, washing hands with soap after
defecation and contact with child stools, and before eating and preparing food stop the
transmission of disease agents because the source of the diarrhea pathogen is removed.
Therefore, washing hands with soap can significantly reduce the burden of diseases associated
with feces and polluted water. The secondary barriers are extremely important when sanitation
services are inadequate and feces are disposed of into the domestic environment.
Figure 4-2: The “F-diagram” shows the pathways of fecal disease transmission and the barriers that can prevent infection
Source: Adapted from Wagner and Lanoix (1959).
4.2.1 Water supply and health
Water is an essential resource needed to sustain life on earth. Numerous empirical studies have
investigated the impact of WATSAN infrastructure on human health. There is compelling
46
empirical evidence that access to improved drinking water supply substantially improves child
health in terms of reducing the risk of diarrheal diseases (Esrey et al. 1990; Fewtrell et al. 2005;
Overbey 2008). Safe, reliable, and easily accessed water is crucial for the preservation of good
health. However, in many developing countries—particularly in rural areas – the available water
is often either unsafe or insufficient to meet basic health needs. The United Nations General
Assembly declared “safe and clean drinking water and sanitation as a human right that is essential
for the full enjoyment of life and all human rights” and openly called for actions leading to the
provision of “safe, clean, accessible and affordable drinking water and sanitation for all” (UN
2010).
The quantity of water available to a given household is largely affected by traveling and waiting
time taken to collect water (Cairncross 1987). After reviewing several studies, Cairncross (1987)
found the following general relationship between water use and collection time: a) the quantity
of water collected decreases considerably once the time taken to collect water is greater than 5
minutes (water source is located 100m away from home); b) the quantity of water remains the
same between 5 and 30 minutes of collection time (distance travelled lies between 100m and
1000m); and c) the quantity of water decreases further when the nearest water source is 1000m
away from home (total collection time longer than 30 minutes). The findings of an empirical study
in Uganda provided evidence to support this observed relationship (WELL 1998). Reducing water
collection time directly increases water availability, and this may translate into more bathing and
washing (Cairncross & Cuff 1987). In particular, the frequency of handwashing is highly correlated
with the quantity of water available to households (Cairncross 1997; Curtis et al. 2000; Gilman et
al. 1993). To reduce the health risk associated with poor hygiene, 50 liters of water per capita per
day (l/p/c/d) is recommended to ensure adequate personal and food hygiene, domestic cleaning,
and laundry needs (WHO 2011); 7.5 liters of water is the recommended absolute minimum for
consumption and cooking in emergencies and disasters (Howard & Bartram 2003). In rural
Ethiopia, the average domestic water consumption is probably much lower than the
recommended amounts, with the estimated amount of drinking water for urban and rural
population varying between 3 and 20 l/p/c/d (Kumie & Ali 2005). This is mainly due to the long
distances between households and their nearest water source, which requires much time and
energy to collect water.
Water access and collection time
In sub-Saharan Africa, Blackden & Wodon (2006) estimated that 40 billion working hours are lost
each year as a result of collecting water. It is also well-documented that it is mostly women and
girls who have to travel long distances and spend their productive time on fetching water
(WHO/UNICEF 2010b). For instance, in Malawi, 87 percent of water fetching duties are taken up
by women and 20 percent of the households spend more than an hour for each water collection
trip (Sorenson et al. 2011). In rural Ethiopia, adult women are 10 times more likely to collect
47
water for household consumption than adult men and 63 percent of the households need to
travel 30 minutes or more for each water collection trip (CSA & ICF International 2012; Usman et
al. 2016). Consequently, it is argued that the time spent collecting water by women and girls
could be used for other income-generating activities, such as seeking access to health care,
schooling, leisure, participating in community activities, and taking care of young children (Ray
2007; Sorenson et al. 2011), with direct and indirect health consequences.
Water quality is also an essential aspect of access to improved water supply. Unsafe drinking
water is not only dirty, but it can also be pathogenic and deadly if people consume it untreated.
More than 1.5 million children die each year due to diarrhea and other gastrointestinal diseases,
mostly in developing countries, and contaminated drinking water is considered to be one of the
major causes (Kremer & Zwane 2007; WHO/UNICEF 2009). Unsafe drinking water may also result
in other waterborne diseases such as typhoid, cholera, and dysentery. The empirical support for
the role of safe water supply in improving and preserving good human health is extensive, in both
rural and urban areas. Most of the existing literature focuses on the impacts of households having
access to piped water connection, and children have been shown to benefit substantially from
improved water sources (Bukenya & Nwokolo 1991; Mangyo 2008). Moreover, the health benefit
of in-house water connection is substantially greater than that of improved public sources such
as standpipe and other protected water sources (Bartram & Cairncross 2010; Curtis et al. 1995).
4.2.2 Sanitation and health
Improved sanitation facility, together with safe water and proper personal hygiene, is
fundamental to good health. In consonance with the World Health Organization (WHO), “no
single type of intervention has a greater overall impact upon national development and public
health than the provision of safe drinking water and the proper disposal of human excreta” (WHO
1996). Lack of adequate sanitation facility can cause diarrheal and other diseases which can be
transmitted via the fecal-oral route. Access to safe and adequate drinking water alone is not
enough to decrease the disease burden because improved sanitation is also a crucial component
of the WASH sector. Improved sanitation technology (e.g., flush toilet and sewerage systems)
creates the primary barriers to prevent fecal pathogens from reaching the environment and can
reduce some health risk factors.
Kumar and Vollmer (2013) analyzed nationally representative data (District Level Household
Survey, DLHS-3) for India and found a 2.2 percentage point reduction in diarrhea incidence in
under-five children living in households with improved sanitation facilities. A similar study in
Nepal by Bose (2009) showed an even larger reduction (11%) among younger children (below 24
months old).
The child stool disposal behavior of households is also an important factor in the prevention of
diarrheal diseases. If children’s feces are not contained or safely disposed of away from the living
48
area, young children might be exposed to the stools through direct contact, which can cause
diarrhea via the hand-to-mouth pathway. Haggerty et al. (1994) found that promoting
handwashing and safe disposal of human and animal excreta from domestic environment results
in an 11 percent reduction in diarrhea morbidity in rural Zaire.
4.2.3 Hygiene behavior and health
Promoting good hygiene behavior is crucial for realizing the full benefit of improved WATSAN
facilities. While WATSAN infrastructure interventions are often considered as the ‘hardware’ of
WASH, the promotion of behavior change is usually considered as the ‘software’ of such
interventions. Hygiene is a set of practice or change of behavior that people adopt to preserve
their health. Changing hygiene behavior is, however, a complex process as it requires people to
change their long-held habits, which have been influenced by cultural and socioeconomic factors.
Although educating people to change their behavior is a complex and uncertain task, Curtis et al.
(2000) suggested that hygiene interventions can be successful if a few behaviors that have most
potential health impacts are targeted and promoted. For instance, promoting handwashing
intervention, and safe water storage and handling practices can produce significant health gains.
Hands are a key vector in the transfer of pathogens from feces as hands can become
contaminated through various mechanisms: during defecation, while disposing of child stool, or
by touching other contaminated objects (Hill et al. 2004). There is, however, a growing body of
evidence that simply washing hands with water and soap at critical times—such as after stool
disposal or defecation, and before preparing food or eating—can help a person avoid life-
threatening water-related diseases. Existing literature also suggested that promoting
handwashing has shown the most success in achieving greater health impact. Curtis and
Cairncross (2003) suggested that handwashing with soap could reduce diarrhea incidence by 47
percent and save one million lives per year. This is consistent with other studies which found that
handwashing interventions achieved a median reduction in diarrhea incidence by 35 percent (Hill
et al. 2004). In a randomized controlled trial in urban Pakistan, Luby et al. (2006) found that
intensive handwashing promotion could reduce diarrhea incidence by 51 percent. Luby et al.
(2005) also analyzed the effect of handwashing with soap on the incidence of pneumonia and
diarrheal diseases and found strong supporting evidence: in households that received the
intervention, diarrhea incidence reduced by 53 percent, and pneumonia incidence by 50 percent
in under-five children. A recent review of hygiene practices by Cairncross et al. (2010) also
indicated that handwashing with soap can result in a 48 percent diarrhea risk reduction in low-
and middle-income countries where there is access to water.
There is also a wealth of evidence indicating that POU water treatment, and safe water storage
and handling lower the risk of exposure to waterborne pathogens, in turn reducing child diarrhea
incidence. A systematic review suggested that treating water with chlorine tablet at the POU
49
reduces not only the risk of Escherichia coli (E.coli) contaminating storage water but also the risk
of child diarrhea significantly in developing countries (Arnold & Colford 2007). Few empirical
studies have investigated the impact of improved water transport and storage containers on
health outcomes. However, a recent study examining the impact of improved water transport
and storage on water quality and health outcomes in Benin by using a randomized control trial
approach found that improved water storage and containers are associated with both a reduction
in E.coli colony count in water and a lower incidence of self-reported diarrheal diseases (Günther
& Schipper 2013).
Improving the quality of drinking water at the household level can bring additional health
improvements. Drinking water collected from improved sources may be contaminated because
of poor water storage and unhygienic water handling before reaching the drinking cup. For
instance, a systematic meta-analysis of 33 studies conducted by Clasen et al. (2007) showed that
water treatment at the POU using flocculation or disinfection is more effective in minimizing the
risk of diarrhea than water source improvements.
Generally, existing empirical studies suggest that some WASH interventions are effective in
reducing water- and fecal-related disease burden. However, the empirical evidence regarding
whether multiple interventions are more effective than single intervention is mixed. Some
studies have shown that using various combinations of interventions are more effective than
using one alone (ALAM et al. 1989; Esrey et al. 1991; van Der Hoek et al. 2001). On the other
hand, a comprehensive meta-analysis by Fewtrell et al. (2005) found that combining
interventions did not have any synergistic effect, which is contrary to the above discussion and
the wider beliefs. Table 4-2 presents the summary of a meta-analysis of the percentage reduction
in diarrheal diseases by intervention type. Waddington et al. (2009) found that water supply
interventions did not bring a statistically significant reduction in diarrheal morbidity but water
quality interventions generated greater diarrheal morbidity reduction. On the other hand,
Fewtrell et al. (2005) and Esrey et al. (1991) respectively found hygiene and sanitation
interventions had greater impacts.
Table 4-2: A meta-analysis of percentage change in diarrheal diseases by intervention type
Esrey et.al. (1991) Fewtrell et.al. (2005) Waddington et.al. (2009)
Water quality -15% -31% -42%
Water supply -20% -25% -2%*
Sanitation -36% -32% -37%
Hygiene -33% -45% -31%
Source: Author’s compilation. *Not significant
50
4.3 Data and Methods
We used the same household survey data collected in two districts of rural Ethiopia between
February and June 2014. It covers 454 households across 20 kebeles with 565 children aged five
or younger. The data comprises household- and community-level data with microbiological water
sample test results.
Microbiological water quality tests
There are several physiological, chemical and microbiological standards for a water supply to be
acceptable for human consumption. One of the most commonly used indicators for microbial
water quality is the level of E.coli bacteria,20 which only comes from human and animal feces. It
is almost impossible to reliably predict the microbial quality of water at the household’s storage
unit based on the type of the water source as water collected from improved sources is often re-
contaminated by fecal matters during collection and transportation due to poor water handling
and storage (Wright et al. 2004). In this analysis, therefore, the level of E.coli bacteria
(CFU/100ml) of storage water sample was used as the indicator of the microbial drinking water
quality (see section 3.3 for a detailed description of the procedure).21
4.3.1 Variables and measurement
a) Dependent variable
In this chapter, we use diarrheal diseases as a health indicator and focused on child health
outcomes.22 In agreement with the WHO, diarrhea is generally defined as the passage of three
or more “loose watery stools” or a loose stool containing blood or mucus in a twenty-four hours
period. Data on the incidence of diarrheal diseases in the two weeks preceding the survey were
collected. In most cases, the respondents are the child’s primary caretaker—usually his/her
mother.
b) Control variables
Based on existing empirical literature, a set of household and child characteristics, water
collection time, latrine characteristics, child stool disposal, and handwashing practices are
included in the analysis to control for observed differences among households. The study,
therefore, makes the hypothesis discussed below concerning the independent effects of
20 We could not test all known pathogens that can pose a health risk because it is both complicated and expensive.
For instance, streptococci and thermotolerant are used as an indicator of drinking water quality as they have a close relationship to bacteria indicators of known fecal origin.
21 The general WHO Drinking Water Quality Guideline suggests that the number of E.coli bacteria CFU/100ml water should be ideally zero when it is used as an indicator of microbiological drinking water quality.
22 Although poor water quality can cause other communicable diseases and can also affect adults, we focused on child diarrheal diseases as the health indicator in this analysis. Young children and infants are more prone to diarrheal diseases than adults due to their weak immune system. Moreover, diarrhea is one of the primary causes of under-five child mortality and morbidity, and contaminated drinking water is considered to be the main cause of diarrheal diseases.
51
explanatory variables on childhood diarrhea. The description of variables and units of
measurement are summarized in Table 4-3.
c) Socioeconomic variables
We start here by considering the “traditional” variables. The ages of both parents and children
are expected to have a negative impact on childhood diarrhea incidence. Younger mothers may
lack the necessary experience to provide better care for their children. Due to their weak immune
systems, infants and young children are susceptible to diarrheal diseases; however, they become
more resistant to the diseases as they grow older. The level of household awareness of the health
benefits of water quality, safe sanitation, and good hygiene practices highly depends on the level
of education among household members and is expected to affect child health positively (i.e.,
decrease diarrhea incidence). As the level of education for both household head and primary
caretaker were extremely low, the highest grade completed among the household members was
used as a proxy for education.
Moreover, in developing countries, socioeconomic factors, such as wealth, also influence the
type of drinking water source used by households (Braind et al. 2010; Larson et al. 2006) and are
expected to be negatively correlated with the incidence of childhood diarrhea. To control for
wealth and other unobserved health practices, household assets value and livestock holding were
used as a proxy for wealth.
A household’s demographic structure may play a role in determining health outcomes.
Household density, dependency ratio, and the number of young children are expected to be
positively associated with childhood diarrhea incidence. In rural areas, housing structures are
poor, with few rooms and crowded living conditions. Consequently, infectious diseases tend to
spread quickly within larger households. However, having a higher proportion of adult women in
a household may reduce childhood diarrhea incidence as children might get better care and more
time, therefore resulting in improved health outcomes.
d) The link between agriculture and WASH
The linkages between WASH, health outcomes and agriculture are crucial. In the context of the
rural households in our study, we looked at two specific areas of such interactions. First, animal
excreta can be a source of pathogens; therefore, keeping livestock may increase the risk of
childhood diarrhea. The sampled households practiced mixed farming, and more than 97 percent
of the households own livestock, which often shares the same space with the people. All these
factors are expected to result in a positive correlation between the presence of livestock and
child diarrhea incidence, although it might be partly offset by the nutritional impacts of a more
diverse (animal protein) diet. Importantly, we recorded not only livestock ownership, which could
be used as a proxy for the wealth effect but also the presence of livestock in/around the human
living area. Second, we tested for the impact of irrigation on childhood diarrhea. The theoretical
relationship between these two factors is unclear. The availability of irrigation water may allow
52
households to use more water than households without irrigation, thus having a positive impact
on sanitation activities and resulting in a lower diarrhea incidence among young children. The
quality of irrigation water may, however, create new sanitation and hygiene issues and therefore
reduce the positive impact of water availability.
Finally, we controlled for basic child health parameters by using exclusive breastfeeding for the
first 6 months of life as an indicator of parental care towards a child’s health, which is strongly
correlated with health outcomes.
e) Water, sanitation, and hygiene
Improved household water quality is expected to reduce the risk of childhood diarrhea by acting
as a barrier to disease-causing pathogens. The traveling and waiting time used for collecting
water determines the amount of water collected by a given household and reduces the time
available for child care and other activities. A recent empirical study has shown that the time
spent on fetching water from distant sources for domestic use significantly affects child health
(Pickering & Davis 2012). Insufficient water may also limit good hygiene practices such as washing
hands regularly at critical times. Overall, water collection time is assumed to be positively
correlated with childhood diarrhea incidence. We also controlled for the practice of handwashing
with soap, which is a defensive mechanism to improve household health and therefore expected
to be negatively correlated with diarrhea incidence.
4.3.2 Descriptive statistics
Table 4-3 shows the descriptive statistics of variables of the sample households used in our
empirical analysis. Some of the variables, which are not shown here, have been already
presented in Table 3-3. As we have discussed earlier in the preceding chapter, the level of
education is exceptionally low in the study areas. Although household heads are much better
than the primary caretakers in terms of self-reported literacy (44% as opposed to 9%). Yet, few
household heads in our sample have completed primary school (less than 4%).
53
Table 4-3: List of variables: Definitions and summary statistics
Variables Description N Mean SD
Household characteristics
Household head literacy 1= read & write; 0=otherwise 454 0.44 0.50
Primary caretaker literacy 1= read & write; 0=otherwise 453 0.09 0.29
Highest education The highest grade completed in a household 454 3.50 3.05
Proportion of adult female Share of female household members aged > 14 454 1.22 0.49
Under 8 years children Number of children under 8 years 454 1.99 0.72
Dependency ratio Share of household members aged below 15 and
above 64 to those aged 15-64
454 0.55 0.12
Consumption expenditure Total household consumption expenditure per adult equivalent in 100 Birr
454 1.46 6.67
Under 5 child characteristics
Age Child age in months 562 29.02 16.30
Gender 1=male; 0=female 565 0.46 0.50
Breastfeeding Exclusive breastfeeding for the first 6 months (1=yes) 562 0.69 0.46
Medical visit Number of medical visits in the past one year 565 0.80 1.28
Child diarrhea Having diarrhea in the last 2 weeks (1=yes) 562 0.16 0.36
Water and sanitation
Drinking water source 1=improved; 0=otherwise 454 0.50 0.50
Water quality 1=uncontaminated; 0=otherwise 454 0.42 0.49
Minutes to water source Time needed for a round trip for water collection 454 20.40 15.72
Time spent fetching water Total time spent in hours over the last 7 days before the survey
454 16.53 13.64
Pit latrine 1=yes; 0=otherwise 454 0.42 0.49
Safe child stool disposal 1=yes; 0=otherwise 454 0.36 0.23
Source: Author’s computation based on own survey data.
Other crucial descriptive statistics in Table 4-3 include WATSAN infrastructure. While 50 percent
of the households in our sample have access to improved water supply (based on the JMP
definitions), about 58 percent of the households had contaminated drinking water at storage.
The results suggest that the JMP definition of ‘improved’ drinking water sources overestimates
access to improved drinking water when taking in consideration POU water safety or quality. The
study also reveals that only 5 percent of the households had access to improved water source in
their own yard or living area, and 34 percent of the households took 30 minutes or more for a
round trip to obtain drinking water (see, Table 3-3). As a result, households spent much time
fetching water, on average about 2 hours and 22 minutes per day.
Improved sanitation facilities are virtually non-existent in rural communities of Ethiopia. About
42 percent of the households reported that they have simple pit latrines while 58 percent of the
households defecate in the open. The reported open defecation rate is much higher than the
rural national average open defecation rate of 43 percent (WHO/UNICEF 2014). Moreover, most
adult women prefer defecating in a bush. The survey also revealed that more than 74 percent of
54
the primary caretakers practiced open defecation for the last stool before the survey. As many
of the latrines were constructed in response to a push by the local government, open defecation
is still a norm and practiced by most people in the study areas. Furthermore, due to the limited
awareness of the harmful nature of child stools, only 33 percent of last child stools preceding the
survey were adequately disposed of.23 As Figure 4-3 shows, about 20 percent of children’s stool
dropped into the toilet or latrine, 2 percent of children’s stool were rinsed or washed away and
water is discharged into the latrine, 11 percent of children’s stool were buried. In contrast, 25
percent of the children’s stools were not removed safely, that is, the stools of a child in every
four are left in the open. Studies have shown that sanitary disposal of fecal matters is an effective
mechanism for reducing child morbidity (Curtis et al. 2000). Leaving child stools in the immediate
vicinity or yard increases the risk of young children coming into direct contact with stools, which
causes diarrheal diseases.
Figure 4-3: Method of disposal for the last child stool preceding the survey
Source: Author’s computation using survey data.
a) Diarrhea as a child health problem
Our illustration shows that 16 percent of under-five children were reported by their primary
caretakers to have experienced diarrheal illness during the two weeks preceding the survey
(Table 4-3). The incidence of child diarrhea in relation to child and parental education
23 A report also shows that there is clear difference between urban and rural areas in the way children’s stools are
disposed of. Nationally, a higher proportion of urban children’s stools (63%) are disposed of safely than of rural children’s stools (31%) (CSA & ICF International 2012).
0 5 10 15 20 25
Did nothing/left it there
Dropped into toilet facility
Disposed outside premises
Buried
Washed away & water discharged outside
Disposed some where in the yard
Do not know
Washed away & water discharged into toilet
Disposed into solid waste/trash
Percentage of children
55
characteristics is presented in Table 4-4. Diarrhea incidence was higher among children younger
than 24 months, particularly among those between 6 and 23 months old. Children aged between
36 and 47 months old were the least affected by diarrhea. However, there was no significant
difference in diarrhea incidence by the sex of the child; boys were slightly more likely to have
experienced diarrhea (16.34%) than girls (15.08%).
Surprisingly, the reported diarrhea incidence was low among the children with illiterate primary
caretakers and household heads. We had expected that diarrhea incidence to decrease with
increasing education levels and literacy of primary caretakers and household heads. This could
have been the result of a lower level of school attainment. For instance, only 4.3 percent of the
household heads and 1.2 percent of the primary caretakers have completed primary school
education. There is not much variation in the outcome of education. Also, this observation could
also be due to primary caretakers who are illiterate having lower knowledge of the symptoms of
childhood diarrhea.
Finally, children living in households with corrugated iron sheet roofing had lower diarrhea
incidence than those living in households with thatch roofing (15.40% and 18.37% respectively).
Although the physical characteristics of a household’s environment often serve as indicators of
the household’s socioeconomic status, they are also considered to be an important determinant
of the health of children and other household members (Cattaneo et al. 2009).
Table 4-4: Diarrhea incidence by demographic characteristics (last two weeks)
Age of child (months) Incidence of diarrhea (%) Number of children
< 6 11.54 52
6 -11 30.36 56
12 -23 33.02 106
24- 35 12.03 133
36 – 47 4.10 122
48 - 59 9.96 93
Sex of child
Male 16.34 257
Female 15.08 305
Primary caretaker’s literacy
Yes 17.39 46
No 15.50 516
Household head’s literacy
Yes 16.36 238
No 14.71 324
Housing roofing material
Corrugated iron sheet 15.40 513
Thatch 18.37 49
Source: Author’s computation using survey data.
56
b) The effect of improved water supply and sanitation on child diarrhea
Diarrhea incidence among under-five children by the household WATSAN characteristics is
presented in Table 4-5. Based on the WHO definition of improved water sources, children living
in households with unimproved water sources were more likely to have had diarrhea in the past
two weeks (18.34%) than those living in households with improved water sources (12.82%). This
result is consistent when broken down by detailed water sources. Children living in households
with private protected water had lower diarrhea incidence compared to those living in
households with shared protected water sources. This might be because water is more available
to households with private protected water, therefore resulting in less interruption to because
of long traveling distance. Moreover, since water is available near to their living area, household
members, especially adult women, can use the time and energy saved to provide better care for
their children. However, diarrhea incidence was much higher in households with contaminated
POU water (24.46%) than a household with uncontaminated POU water.
The hygiene habits of a child’s caretaker were also an important factor. As shown in Table 4-5,
the diarrhea incidence among children was lower if their caretaker practiced handwashing
(18.02% as opposed to 9.55%). Diarrhea incidence was also higher among children whose
households treated their drinking water, although the number of such households is small.
Households that treated their water likely have water sources with poorer quality than
households that do not treat water. However, children in households treated their water had
lower diarrhea incidence than those in households that did not treat their water. Households in
the study areas seldom practiced water treatment. Slightly below 8 percent of households
reported that they use some form of water treatment with 81 percent of these households
applying chlorine-based methods to treat their water during the month preceding the survey. On
the other hand, no significant difference in diarrhea incidence was observed between irrigating
and non-irrigating households in the two weeks preceding the survey. There was also no
significant difference in diarrhea incidence between households with and without a pit latrine.
57
Table 4-5: Diarrhea incidence by WASH characteristics (last two weeks)
Incidence of diarrhea (%) Number of children
Water source based on WHO
Improved source 12.82 273
Unimproved source 18.34 289
Water sources
Private protected dug well 6.67 30
Shared protected dug well/spring 13.58 243
Unprotected well/spring 18.83 223
Surface water 16.67 66
Storage water quality
Contaminated 24.46 327
Uncontaminated 3.4 235
Handwashing with soap
Yes 9.55 157
No 18.02 405
Water purification/treatment
Yes 5.00 40
No 16.48 522
Latrine
Yes 15.41 231
No 16.02 331
Irrigation farming
Yes 15.38 377
No 16.22 185
Source: Author’s computation using survey data
c) Health knowledge and hygiene awareness
While 77 percent of the primary caretakers thought that diarrhea can be prevented, most did not
see that poor water quality and lack of proper hygiene and sanitation as potential causes of
childhood diarrhea. In most cases, the primary caretakers considered contaminated food as the
major cause of diarrhea in young children, followed by bad or poor water quality. While 15.42
percent of the primary caretakers did not know what causes diarrhea, other factors, such as poor
hygiene, dirty hands, flies and germs, and poor sanitation practices, were cited as causes of high
diarrhea incidence by the study participants (Figure 4-4).24
24 A report from the Mecha woreda health center, one of the study district, showed that pneumonia, diarrhea (non-
bloody), malaria PF, acute febrile illness (AFI), acute upper respiratory infection, malaria PV and infection of skin were the top seven diseases morbidity for under-five years in 2013.
58
Figure 4-4: Causes of childhood diarrhea reported by the primary caretakers
Source: Author’s computation using survey data.
4.4 Empirical Strategy
This section outlines the empirical strategy employed to explain the impact of household drinking
water quality and sanitation behavior on child health outcomes. The dependent variable
indicates the self-reported diarrhea incidence in the two weeks preceding the survey. To identify
a robust relationship between POU water quality and sanitation and health outcomes, the
robustness of the results was examined using a range of estimation strategies.
Establishing the causal impact of drinking water quality on child health outcomes based on cross-
sectional data is difficult as it requires a careful investigation of the treatment variable to address
the possible endogeneity problem. For instance, endogeneity can arise where there is an
unobserved covariate that determines both water quality and health outcomes.25 In this analysis,
POU drinking water quality was used as the treatment variable. Water quality was determined
based on the results of the microbiological water sample tests during the data collection period.
In line with the WHO guidelines for drinking water quality, water is considered unsafe or
contaminated if the E.coli colony count per 100ml of water is greater or equal to one. In this case,
however, endogeneity of sanitation variables was not an issue as we used village-level
(neighborhood mean) indicators for both pit latrine and the disposal of child stools.
25 Moreover, the water quality variable may be an endogenous regressor due to unobservable heterogeneity among
household members or omitted variables which cannot be captured in our data affecting both household drinking water quality and health or measurement errors.
293
227
129110 108
7033 12
0
50
100
150
200
250
300
bad/dirtyfood
bad/dirtywater
flies dirty hands poorhygiene
do notknow
feces oropen
defecation
germs
Nu
mb
er o
f h
ou
seh
old
s
59
4.4.1 Instrumental variable approach
Given the likely endogeneity of POU drinking water quality, inferring the causal impact of water
quality on health outcomes from cross-sectional data is difficult. In this chapter, the impact of
drinking water quality on child health outcomes was analyzed using a two-stage instrumental
variables approach. For the first stage consider the following linear probability model:
𝑊𝑖𝑗 = 𝑋𝑖𝑗𝛽1 + 𝑍𝑖𝑗𝜂1 + 𝑁𝑗𝜇1 + 𝜐𝑖𝑗 (4.1)
where treatment 𝑊𝑖𝑗of each household 𝑖 in community j is predicted using a vector of household
characteristics 𝑋𝑖𝑗; 𝑍𝑖𝑗 is a vector of instrumental variables; 𝑁𝑗 is a vector of socio-demographic
factors, which is constant within a community j; and a non-systematic error term 𝜐𝑖𝑗, which varies
over households such that 𝐸[𝑣𝑖𝑗|𝑋𝑖𝑗, 𝑍𝑖𝑗 , 𝑁𝑖𝑗] = 0. While 𝛽1, 𝜂1 and 𝜇1 are unknown parameters
to be estimated in the first stage. The second stage employs the predicted treatment status 𝑊𝑖�̂�
from Equation (4.1) to estimate the treatment effect on outcome𝐻𝑖𝑗, such that
𝐻𝑖𝑗𝑘 = 𝑋𝑖𝑗𝛽2 +𝑊𝑖�̂�𝜃 + 𝑁𝑗𝜇2 + 휀𝑖𝑗 (4.2)
where 𝐻𝑖𝑗𝑘 denotes the outcome (e.g., diarrhea) for child 𝑘 in household 𝑖 and in community 𝑗,
𝑋𝑖𝑗 is a vector of household- and child-specific characteristics, and 𝑁𝑗 are the same covariates as
used in stage 1.
4.4.2 Bivariate probit estimator
As an alternative to the standard linear instrumental variable (IV) methods, Greene (2012) has
shown that average treatment effects (ATE) can be obtained by a bivariate probit model (BP).
The BP model is a two-equation binary outcome model with correlated error disturbances. The
disturbance terms of the two equations are assumed to be jointly distributed as standard
bivariate normal. In this approach, the models are estimated simultaneously using maximum-
likelihood (ML) estimation.
To account for any potential selection effects, the endogeneity of treatment (𝑊) and outcome
(𝐻) may be modelled jointly based on the assumption that the treatment has a direct causal
impact on the outcome and both are influenced by common observable factors. Suppose 𝐻
represents the observed health status of a child and takes a value of 1 if a child had diarrhea in
the last two weeks preceding the survey, and zero if otherwise, then the observed response
variable 𝐻is related to an unobserved latent variable 𝐻∗ as follows:
60
𝐻𝑖𝑗𝑘∗ = 𝛼1𝑖𝑋𝑖𝑗 + 𝛼2𝑖𝑊𝑖𝑗 + 휀1𝑖
𝐻𝑖𝑗𝑘 = {1𝑖𝑓𝐻𝑖𝑗𝑘
∗ > 0
0𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
(4.3)
where 𝑋 is a vector of control variables; 𝑊 is a dummy variable with a value of 1 if the water is
uncontaminated, and zero if otherwise; 𝛼1is a vector of parameters to be estimated; 𝛼2is the
parameter of interest associated with the dummy variable; 휀1𝑖 is a residual term, with𝐸[휀1𝑖] = 0
and 𝑣𝑎𝑟[휀1𝑖] = 1; and 𝑖 indexes households.
Equation (4.3 gives an unbiased parameter estimate of 𝛼 based on the assumption that 𝑊 is an
exogenous variable. But the validity of this assumption can be questioned because the observed
variation in 𝑊may reflect the unobserved factors that also influence the outcome variable𝐻.
Within a given neighborhood, some households may have unobserved preferences that causes
them to have better household water quality than other similar households (Koolwal & Van de
Walle 2010). Therefore, a simplistic comparison of child health status between households with
contaminated water and households with uncontaminated water would lead to biased results.
To account for the endogeneity of the water quality variable𝑊, the bivariate probit model was
constructed as:
𝐻𝑖𝑗𝑘∗ = 𝛼1𝑖𝑋𝑖𝑗 + 𝛼2𝑖𝑊𝑖𝑗 + 휀1𝑖
𝑊𝑖𝑗∗ = 𝛽1𝑖𝑋𝑖𝑗 + 𝛽2𝑖𝑍𝑖𝑗 + 휀2𝑖
𝑊𝑖𝑗 = {1𝑖𝑓𝑊𝑖𝑗
∗ > 0
0𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
(4.4)
where 𝑊𝑖𝑗∗ represents a latent continuous variable, 𝑍 is a vector of instrumental variables, and by
assumption we have 𝐸⌈휀1𝑖⌉ = 𝐸⌈휀2𝑖⌉ = 0 and 𝑣𝑎𝑟⌈휀1𝑖⌉ = 𝑣𝑎𝑟⌈휀2𝑖⌉ = 1 with 𝑐𝑜𝑣⌈휀1𝑖휀2𝑖⌉ = 𝜌.
Rho measures the correlation between omitted or unobserved factors in the health and water
quality equations (Wooldridge 2010). 𝛽1and 𝛽2 are vectors of parameters to be estimated, and
a Wald test can be applied to 𝜌 to test for the exogeneity of𝑊. If 𝜌 is significantly different from
zero, then we can estimate the two equations jointly with a bivariate probit approach using
maximum likelihood (ML), and it will produce consistent estimates. Here, we estimated a
recursive bivariate probit model whereby water quality appeared as a regressor in the health
outcome equation.26
26 Generally, Nichols (2011) showed that in case of a binary regression with a binary endogenous variable linear IV
generates robust consistent estimates of the ATT (average treatment on the treated) while bivariate probit produces efficient estimates of the ATE.
61
4.5 Estimation Results and Discussion
To address the possible endogeneity of POU water quality, we proposed two variables as
instruments for the treatment variable. The first instrument is the type of primary sources from
which households fetch their drinking water. In the study areas, most households take water
from community water sources, and most rural households do not have alternative drinking
water sources. In our sample, only 5 percent of the households have access to an improved
private water source in their living area. The second instrument is the existence of water user
association/committee in the village. We asked households whether there is a WUA responsible
for taking care of community water sources in the village. Water user associations are primarily
responsible for monitoring, supervising and handling conflicts among household users of
community water sources. Such groups are instituted in many villages for governing rural
communal water sources when a new water source is developed. For instance, 62 percent of the
villages in Mecha district had water user groups (Tilahun et al. 2013). Water source and water
user group can both be treated as exogenous variables in the context of this study as they do not
affect child health outcomes directly.
The impact of safe water and sanitation behavior on childhood diarrhea estimation is presented
in Table 4-7. Columns 1 to 3 report the linear probability model (OLS) and the binary probability
model (Probit), which do not take into account any endogeneity problem. Estimating the “naïve”
model helps to examine the extent to which our results are sensitive to the assumption that
stored household water quality is an exogenous variable. Columns 4 to 7 show the standard
instrumental variable (IV) models and the recursive bivariate probit (BP) models. As the
regression results show, drinking water quality, safe child stool disposal and latrine density
significantly affected child diarrhea incidence in all model specifications. The estimates of other
relevant variables coefficients are also statistically significant with the expected signs across all
specifications.
Both the linear IV and BP two-stage models require a strong treatment prediction in the first
stage. The first-stage regressions showed a statistically significant relationship between the
treatment and instrument variables in Table 4-6. This relationship is robust with and without
second-stage controls. The r-squared is modest, and a large F-statistics in the first-stage
regression suggests bias from weak instruments is unlikely to be a problem. For the linear IV
model, the over-identification restriction test regarding the instruments is not violated (p=0.764),
which implies that we can reject the null hypothesis that at least one of the instruments is invalid.
On the other hand, considering the exogeneity test for household water quality, Wooldridge’s
score (robust regression) test does not reject the null hypothesis that water quality is exogenous
at any of the conventional significance levels (p=0.867), as presented in Table 4-6. Even if stored
household water quality were exogenous, the linear IV estimates are still consistent but are less
efficient than the least-square estimates.
62
Table 4-6: Household water quality: First-stage regression
(1) (2) (3) (4) VARIABLES LEAST SQUARES PROBIT
Water source (1=protected) 0.134** 0.124** 0.127*** 0.115*** (0.055) (0.048) (0.049) (0.042) Water user association 0.383*** 0.328*** 0.337*** 0.273*** (0.060) (0.058) (0.048) (0.046)
Observations 565 562 565 562 Stage 2 controls NO YES NO YES R-squared 0.19 0.32 Model F-Test 53.37 21.44 Model Chi2 78.03 186.09 Instruments jointly p-value 0.000 0.000 0.000 0.000 Basmann over-identification p-value 0.764 0.915 Robust regression test for endogeneity p value 0.867 0.958
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1 Probit in average marginal effects.
Moreover, the exogeneity test from the BP framework, as shown in columns 6 and 7 (Table 4-7),
indicated that 𝜌̂ is not statistically different from zero (p=0.648). The result indicated that the
error terms are independent, that is, the variable stored water quality can be treated as
exogenous. The BP model is, therefore, equivalent to two independent probit models.
In the preferred probit model (column 2 of Table 4-7),27 uncontaminated stored water and safe
child stool disposal decreased the incidence of child diarrhea, whereas a higher pit latrine density
increased the risk of diarrhea for under-five children in all model specifications. The impact of
safe drinking water on child diarrhea incidence was modest and statistically significant at 1
percent level, with a marginal effect of 0.15; that is, the probability of child diarrhea was 15
percentage points lower in households with safe household drinking water. Safe child stool
disposal decreased child diarrhea incidence by 23 percent, and pit latrine density increased it by
13 percent. The finding that neighborhood pit latrine concentration increased the risk of diarrhea
in young children casts serious doubt on the assumed health and social benefits of moving from
open to fixed-location defecation.28
27 The probit regression model is preferred because it is more efficient than the 2SLS or BP when endogeneity is not
a problem. 28 This indicates that household level variation alone does not seem to be sufficient to capture children’s potential
exposure to poor WATSAN. Rather, the type of sanitation technologies and the practice of child stools disposal at your neighbour are also important determinants of a child’s health status.
63
Table 4-7: Health effects of water quality and sanitation: Diarrhea in under-five children
(1) (2) (3) (4) (5) (6) (7) VARIABLES OLS Probit Probit 2SLS 2SLS BP BP
Water quality (1= no E.coli) -0.140*** -0.149*** -0.148*** -0.119** -0.137** -0.119* -0.125* (0.027) (0.031) (0.030) (0.059) (0.058) (0.070) (0.073) Child age in months -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Child is male 0.005 0.003 0.002 0.005 0.007 0.003 0.002 (0.026) (0.021) (0.021) (0.025) (0.025) (0.021) (0.021) Mother age -0.062*** -0.052*** -0.049*** -0.062*** -0.058*** -0.052*** -0.048*** (0.023) (0.017) (0.018) (0.022) (0.022) (0.017) (0.017) Mother age square 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Highest education completed 0.003 -0.000 0.002 0.002 0.006 -0.001 0.001 (0.006) (0.005) (0.005) (0.006) (0.006) (0.005) (0.006) Number of adult female -0.073** -0.073* -0.068* -0.076** -0.072** -0.075** -0.070** (0.030) (0.037) (0.036) (0.031) (0.031) (0.036) (0.035) Household density 0.046** 0.036*** 0.037*** 0.047*** 0.048*** 0.037*** 0.038*** (0.019) (0.012) (0.012) (0.018) (0.018) (0.012) (0.012) Exclusive breastfeeding (dummy) -0.083*** -0.071*** -0.070*** -0.083*** -0.084*** -0.070*** -0.070*** (0.026) (0.021) (0.021) (0.026) (0.027) (0.021) (0.021) Dependency ratio 0.024 0.040 0.068 0.016 0.073 0.031 0.061 (0.150) (0.131) (0.135) (0.139) (0.146) (0.126) (0.131) Number of medical visits 0.043*** 0.032*** 0.032*** 0.043*** 0.043*** 0.032*** 0.032*** (0.014) (0.009) (0.009) (0.013) (0.013) (0.009) (0.009) Minutes to water source 0.002* 0.002** 0.002** 0.002* 0.002* 0.002** 0.002** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Handwashing with soap (dummy) -0.055* -0.059* -0.059* -0.058* -0.059* -0.062* -0.061* (0.032) (0.033) (0.032) (0.033) (0.032) (0.035) (0.034) Livestock holding 0.021** 0.020** 0.022** 0.023** 0.025** 0.023** 0.024** (0.010) (0.010) (0.010) (0.011) (0.010) (0.011) (0.011) Irrigating households (dummy) -0.005 -0.009 0.003 -0.004 0.014 -0.007 0.003 (0.036) (0.037) (0.042) (0.035) (0.039) (0.037) (0.041) Log of assets value -0.046** -0.037** -0.030* -0.047** -0.034* -0.039** -0.031* (0.020) (0.017) (0.018) (0.019) (0.020) (0.017) (0.017) Safe child stool disposal -0.216*** -0.229*** -0.226*** -0.218*** -0.219*** -0.229*** -0.225*** (0.078) (0.084) (0.086) (0.076) (0.077) (0.083) (0.085) Latrine density 0.141** 0.135** 0.128** 0.137** 0.129** 0.129** 0.124** (0.066) (0.062) (0.058) (0.064) (0.062) (0.063) (0.060) Log of total expenditure -0.051 -0.086** -0.051 (0.041) (0.037) (0.041) Distance to health center 0.000 0.000 0.000 (0.004) (0.004) (0.004) Constant 1.467*** 1.468*** 1.857*** (0.304) (0.296) (0.344)
Observations 562 562 562 562 562 562 562 Model F-Test 14.58 Model Chi2 232.22 252.73 273.79 295.08 652.61 809.36 Model p-value 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Probit rho Chi2 0.208 0.108 Probit rho p-value 0.648 0.742
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1; and Probit and BP in average marginal effects.
64
Regarding the control variables, most of the estimated coefficients had the expected signs. Child
age is significant and negative, implying a reduction of diarrhea incidence as a child grows older.29
In terms of gender, male and female children are equally affected by diarrhea. The relationship
between a mother’s age and diarrhea incidence in her children is nonlinear, and this indicated
that children with a younger mother tend to fall ill more often. However, the age of a household
head had no significant impact on childhood diarrhea incidence. Also, the level of household
education showed no clear effect. This is because usually the primary caretaker’s level of
education is most important in improving child health.
Time to a water source was marginally significant at 5 percent level, but the estimated effect on
diarrhea incidence was much smaller than some of the estimates reported in the existing
literature in sub-Saharan Africa (Pickering & Davis 2012). The estimated coefficient is extremely
small, and this might be because households with water sources on their premises are not better
off due to contamination from own latrine and unhygienic water withdraw from sources.
Distance to the water source also affected the quality and quantity of water a household could
collect. The farther a water source from home, the less water a household could collect and use.
In addition to affecting health, inadequate water quantity may cause households to limit their
handwashing and key hygiene practices and to wash dishes and clothes using water from
unimproved sources. This evidence further suggested that the health benefits of having access
to improved water may not be substantial if water sources are further away from the house and
therefore requiring much time and energy to fetch the water. Moreover, closer proximity to a
water source could have an indirect health benefit. By reducing the time spent on collecting
water, more time is available for looking after children or engaging in other productive and
income-generating activities. In terms of hygiene behaviors, we also found that handwashing
with soap marginally improves child health outcomes. Handwashing with soap is considered to
be an effective defensive mechanism to remove germs and pathogens from hands.
Children belonging to a household with a greater number of adult women seemed to be better
off Table 4-7. It can be argued that children with an additional adult woman get better care. The
household density variable had a positive sign and was statistically significant at 1 percent level.
It was expected that the relative disease burden would increase in households living in congested
or overcrowded conditions. It seemed plausible that children living under such conditions are
more exposed to health risks. Moreover, children in larger families are more exposed to health
risks because of the larger number of hand-to-water pathogen transmission pathways. On the
other hand, household dependency ratio did not have any significant effect on child health.
Exclusive breastfeeding also had a significant effect (7%) on diarrhea incidence in young children.
It is one of the most robust findings in determining childhood diarrhea with large protective
29 Child age-squared term was excluded in the regression as it is not statistically significant, but children between 6
and 24 months old were the most affected age group in this study.
65
effect. We expected a much stronger effect for children under-three years than under-five years
but the impact is the same. However, this variable is more likely to suffer from recall bias or
measurement error. The number of medical visits in the previous year significantly increased with
diarrhea incidence. This variable is used as a proxy for the general health status of a child in the
previous year. We expected that children who often fall sick have a weak immune system and
are therefore more susceptible to diarrheal diseases.
We could not find any significant difference in childhood diarrhea incidence between irrigating
and non-irrigating households. However, livestock ownership increased diarrhea incidence, and
its coefficient was statistically significant at 5 percent level. On the other hand, we expected
children from wealthier households to be better off than those from poorer households because
higher income allows a household to access better health-care services and invest in measures
that improve their health. Though the estimated effect is marginally statistically significant, its
economic relevance is quite small as the value is logged. Distance to the nearest health center
did not have any significant impact on health either. It is evident from this study that stored
drinking water quality, sanitation, handwashing and breastfeeding habits are important factors
in determining the risk of childhood diarrhea.
Our findings regarding the health effects of latrine contradict the claim that moving from open
defecation to the first ladder of sanitation services generates the greatest amount of benefits
from sanitation services. There could be a few reasons for this apparent paradox: first, pit latrine
attracts flies, which are vectors of pathogens that can transmit diseases through direct contact
with young children or food; second, we found that availability of latrine deteriorates POU water
quality, particularly for households whose water sources are located within the living area. The
sub-sample analysis presented in the appendix Table A1 showed that the concentration of E.coli
is significantly higher in households whose water source was located near their living area, even
after controlling for other household characteristics (p=0.001). Well water may be polluted by
leachates from latrines if the latrines are not located sufficiently far away from water sources.
Other studies have shown that in rural areas, latrines can be a source of ground water pollution
(for a detail discussion see Banks et al. 2002; Graham & Polizzotto 2013). Poorly maintained
latrines can generate negative externalities, which affect not only its owner but also the
neighboring communities. There must be a shift in government policy away from building simple
pit latrines to create open defecation free villages in rural areas because it may not generate the
desired health benefits from sanitation services. Existing latrines should be upgraded to make
them safer and hygienic as simply adopting pit latrines may worsen a community’s health status
at large.
Generally, the study highlights the evidence that stored drinking water quality and sanitation
issues are a great public health concern in rural Ethiopia. Infants and young children are more
likely to suffer from water- and fecal-related diseases than any other age groups. Promoting the
66
following among the rural communities could substantially reduce the transmission of germs and
pathogens that can cause diarrheal and other infectious diseases: 1) washing the hands of
caretakers and children with soap at critical times, 2) adequate household water treatment and
safe handling of water, 3) safe disposal of child stools, and 4) good hygiene practices when
feeding and handling food. On the one hand, women are primarily responsible for collecting
water for household use and other household chores, and they are usually the ones reinforcing
hygiene practice at home. On the other hand, the level of education in the study areas is
extremely low, especially for mothers (primary caretakers). Primary caretakers and women are
therefore an important target group of any WASH-related interventions for them to be effective
in reducing overall diseases burden.
Robustness check
The impacts of drinking water quality, safe stool disposal, and latrines on child diarrhea incidence
are robust to model specifications and age. The result of the subsample analysis for under-three
children (presented in Table 4-8) showed that good water quality and proper child stool disposal
have a larger positive impact on younger children, while the impact of latrine density remained
the same. Moreover, almost all the results are statistically not different from one another. With
the regression estimates shown in (column 1, Table 4-9), we measured the effects of latrines at
the household level, and the results showed that latrines had still a strong positive impact (8%)
on child diarrhea incidence. However, the effect of water quality on child diarrhea is not big when
E.coli coliform counts are considered as a continuous variable (presented in column 2, Table 4-9).
Moreover, the way we defined ‘safe child stool disposal’ created some correlation between the
variables ‘safe child stool disposal’ and ‘latrine’. However, it did not cause much problem because
less than 22 percent of the households used a toilet to dispose of the child stool.
67
Table 4-8: Health effects of water quality and sanitation: Diarrhea in under-three children
(1) (2) VARIABLES PROBIT SE PROBIT SE
Water quality (1= no E.coli) -0.199*** 0.044 -0.199*** 0.043 Child age in months -0.005*** 0.002 -0.005*** 0.002 Child is male 0.010 0.031 0.009 0.031 Mother age -0.060** 0.024 -0.057** 0.024 Mother age squared 0.001*** 0.000 0.001** 0.000 Highest education completed 0.002 0.007 0.003 0.007 Number of adult female -0.082* 0.049 -0.081* 0.048 Household density 0.039** 0.016 0.039** 0.016 Exclusive breastfeeding (dummy) -0.074** 0.037 -0.074** 0.037 Dependency ratio 0.025 0.178 0.034 0.183 Number of medical visits 0.041*** 0.012 0.041*** 0.012 Minutes to water source 0.002** 0.001 0.002** 0.001 Handwashing with soap -0.056 0.043 -0.055 0.043 Livestock holding 0.014 0.015 0.015 0.015 Irrigating households (dummy) -0.020 0.048 -0.017 0.054 Log of assets value -0.046* 0.026 -0.043 0.027 Safe child stool disposal (village mean) -0.324*** 0.103 -0.307*** 0.111 Latrine density (village mean) 0.137** 0.065 0.138** 0.068 Log of total household expenditure per adult equiv. -0.021 0.056 Distance to health center 0.002 0.006
Observations 361 361 Pseudo R-squared 0.30 0.30
Model Chi2 154.98 206.70
Model p-value 0.000 0.000
Robust standard errors adjusted for clustering at village level; Significance *** p<0.01, ** p<0.05, * p<0.1 Probit in average marginal effects.
68
Table 4-9: Health effects of water quality and sanitation: Diarrhea in under-five children
(1) (2) VARIABLES PROBIT SE OLS SE
Contaminated water quality (log(1+E.coli) 0.082*** 0.008 Water quality (1=no E.coli) -0.150*** 0.033 Pit latrine (dummy) 0.078** 0.035 Latrine density (village mean) 0.157** 0.060 Safe child stool disposal -0.186** 0.078 -0.201** 0.083 Child age in months -0.004*** 0.001 -0.004*** 0.001 Child is male 0.005 0.020 0.009 0.025 Mother age -0.055*** 0.018 -0.049** 0.021 Mother age squared 0.001*** 0.000 0.001** 0.000 Highest education completed 0.000 0.005 0.003 0.005 Number of adult female -0.079** 0.038 -0.050 0.031 Household density 0.038*** 0.012 0.028* 0.016 Exclusive breastfeeding (dummy) -0.071*** 0.022 -0.078*** 0.025 Dependency ratio 0.077 0.134 0.002 0.142 Number of medical visits 0.032*** 0.009 0.036*** 0.013 Minutes to water source 0.002** 0.001 0.001 0.001 Handwashing with soap -0.063* 0.034 -0.030 0.029 Livestock holding 0.018* 0.010 0.004 0.009 Irrigating households (dummy) -0.013 0.040 -0.023 0.037 Log of assets value -0.040** 0.018 -0.024 0.018 Distance to health center -0.001 0.005 -0.001 0.004 Constant 0.966*** 0.287
Observations 562 562 Model Chi2 218.48 Model F-stat 18.40 Model p-value 0.000 0.000 Pseudo / R-squared 0.33 0.35
Robust standard errors adjusted for clustering at village level; Significance *** p<0.01, ** p<0.05, * p<0.1 Probit in average marginal effects.
69
Longitudinal diarrhea prevalence
It is understood that ‘period prevalence’ is more preferred to ‘point prevalence’ in many health
outcome studies—particularly infectious diseases.30 Morris et al. (1996) argue that longitudinal
prevalence (LP) of diarrhea is more closely associated with child growth faltering and mortality
than the incidence of diarrhea. Therefore, in this section, we measure diarrhea using longitudinal
prevalence rather than incidence. LP can be calculated as the total number of days a child has
diarrhea divided by a total number of days of observation (Morris et al. 1996). We examined what
determines LP for under-five children; is there any significance difference in the LP of diarrhea
between children from irrigator and non-irrigator households. Further, we also discussed the
number of episodes of diarrhea and duration.
The data comes from the same households. Field workers visited each household biweekly, for
12 weeks (April 2014—June 2014), that is, each household was visited a maximum of six times,
and asked the child’s primarily caretaker—usually the mother, if the children had diarrhea
defined as ‘three or more loose stools or a loose stool containing blood or mucus in a twenty-
four hours period’ in the preceding two weeks. If so, the primary caretaker’s also asked for how
many days the symptoms stayed.
Descriptive analysis
The proportion of under-five children with diarrhea in the last two-weeks preceding the survey
in each visit round is presented in Figure 4-5. The minimum prevalence rate was observed during
the 1st visit-round while the maximum was observed during the 5th round (15.96% versus 20.11%,
respectively). The overall average prevalence rate during the whole period was 17.85 percent,
which is slightly higher compared to the baseline prevalence rate of 16 percent. As can be seen
from the figure, diarrhea prevalence rate is, however, drastically increased from 15.96 percent
to 18.94 percent (visit 1 to visit 2) and from 16.7 percent to 20.11 percent (visit 4 to visit 5).
Moreover, based on irrigation status, we find that children from irrigator households are more
likely to experience diarrhea than children from non-irrigator households.
30 Point prevalence is a single assessment of a fixed time whereas period prevalence is defined as “the percentage
of a population that are cases at any time within a stated period.” In period prevalence subjects being followed or observed repeatedly over a period of time (for instance, multiple episodes of diarrhea or other infectious diseases).
70
Figure 4-5: Prevalence of diarrhea for under-five children in the follow-up survey
Source: Author’s computation using own survey data.
The proportion of children with diarrhea by the number of episode is presented in Figure 4-6. It
shows that about 35 percent of the children did not have diarrhea, and one-third of the children
had diarrhea once over the observation period. Moreover, few children had experienced diarrhea
more than three times over the period. On the other hand, the proportion of children with
diarrhea by the number of episodes and irrigation status is presented in Figure A1 (see Appendix
4). The figure suggests that on average children living in irrigator households are slightly more
likely to experience more diarrhea episodes than children living in non-irrigator households (5.22
episodes per child-year compared with 4.35 episodes per child-year). Confirming to the recent
estimates of Walker et al. (2012), the incidence of diarrhea per child year in Africa is 3.4 (2.2, 5.1)
episodes.
15.96
18.9417.38
16.7
20.11
18
5
8
11
14
17
20
23
1 2 3 4 5 6
% o
f ch
ildre
n w
ith
dia
rrh
ea
Number of visits
71
Figure 4-6: Proportion of children with diarrhea by number of episodes
Source: Author’s computation using own survey data.
The proportion of children with diarrhea by the number of episode duration day is presented in
Figure 4-7. It suggests that diarrheal illness lasts for three days for more than one-third of the
children who had already diarrhea over the period. The mean episodes duration of diarrhea is
about 3.4 days; however, as the figure shows few children had persistent diarrhea during the
period (lasting for 10 or more days). Similarly, the proportion of children with diarrhea by episode
duration days and irrigation status is presented in Figure A2 (see Appendix 4). In general, the
results may be suffered from long recall periods (last 14 days).
0
0.1
0.2
0.3
0.4
0 1 2 3 4 5 6
Pro
po
rtio
n o
f ch
ildre
n w
ith
dia
rrh
ea
Number of episodes
72
Figure 4-7: Proportion of children with diarrhea by episode duration
Source: Author’s computation using own survey data.
4.6 Conclusions and Recommendations
Using a combination of estimation methods, this study examined the impact of household
drinking water quality and sanitation behavior on child health in rural Ethiopia. Diarrhea was used
as a health outcome. In this chapter, we focused on children under the age of five because
children in this age group are the most vulnerable to water- and fecal-related diseases. The study
relied on primary survey data from rural areas in Fogera and Mecha districts of Ethiopia. As access
to clean water and improved sanitation is lacking, water-related diseases are the most prevalent
health problem in the study areas.
Some studies have been conducted to quantify the impact of safe drinking water on child health.
However, still little is known about it in the context of rural areas, where access to improved
water is inadequate and the majority of the population rely on unimproved water sources. One
of the innovations of this study is that household drinking water quality was determined by
testing the microbiological quality of household storage water using membrane filtration method
rather than looking at the types of household drinking water sources.
Inadequate access to improved WATSAN facilities remains a major cause of health problems in
Ethiopia, particularly in the rural areas, where a lack of clean drinking water and unsafe sanitation
practices are the main causes of diarrheal diseases among under-five children (CSA & ORC Macro
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
1 2 3 4 5 6 7 8 9 >=10
Pro
po
rtio
n o
f ch
ildre
n w
ith
dia
rrh
ea
Episodes duration (days)
73
2006). The negative health impact of contaminated water is worsening because most rural
households only have access to drinking water from unprotected sources and they often
consume the water without any in-house treatment. Moreover, most rural population have a
poor understanding of the importance of proper hygiene practices, which further increases their
already high risk of contracting infectious diseases.
The findings suggest that access to an improved drinking water source is low in the study areas
—only 50 percent of the households have access to improved water sources. The household
water sample test also indicated that poor POU water quality is a significant problem in rural
Ethiopia. Besides, most of the so-called ‘improved’ water sources in rural areas do not guarantee
the water is safe for consumption, the problem of unsafe drinking water is exacerbated by POU
water contamination through unsafe water storage and handling practices. Even though access
to clean water and simple pit latrines has increased significantly during the last decade, many of
the surveyed rural residents did not regard the progress as satisfactory in terms of access to clean
water supply. Rural households complained about a lack of access to safe water sources, poor
water quality, and having to travel long distances to access drinking water.
In terms of sanitation, we found that 42 percent the households were equipped with a simple pit
latrine while the rest of our sample households defecated in the open. Access to improved
sanitation facility is virtually non-existent in the study areas, and the existing sanitation
technology used there is considered unimproved based on the commonly used WHO definition.
In some cases, these latrines do not have proper structure and become dysfunctional for many
reasons, including the fact that they are not connected to any sewerage system.
Surprisingly, neighborhood latrine concentration increased the risk of childhood diarrhea. We
found that children living in neighborhoods with high latrine density are at higher risk of
contracting diarrheal disease. This indicates that existing pit latrines are not safe to use and do
not protect against diarrheal diseases. Others have argued that the greatest benefit of sanitation,
for both health and social reasons, can be achieved when people move from open to fixed-
location defecation (Mara et al. 2010). However, contrary to the belief of many, our study
suggests that in a rural setting, where settlements are scattered, defecating in the open might
not be more harmful than using a simple pit latrine. Nevertheless, we are not encouraging open
defecation per se, but rather arguing that simple pit latrines are not good enough to achieve the
desired health benefits of sanitation. A study conducted in Ethiopia also found using pit latrines
offered no improvement over defecating in the open in terms of health outcomes (Cameron
2009).
Some policy recommendations can be derived from these findings. First, more efforts should be
put into increasing the existing coverage of improved rural water supply. This can be achieved by
developing new water points and upgrading existing unimproved sources. Improving access to
clean water supply not only increases the quantity of clean water available for household
74
consumption but also allows households to save much time by reducing the distance between
each household and the nearest water access point. Second, pit latrines should only be adopted
if adequate hygiene can be maintained, otherwise, they can pose a serious health risk, especially
if they are not fly-proof or insufficiently away from water wells. Third, a water quality monitoring
system which monitors a set of common water quality indicators should be in place to ensure
rural water supply schemes comply with quality standards. Fourth, household water treatment
and safe water storage should be promoted to address POU water quality concerns. Increasing
the provision of rural water supply alone may not be enough if households do not treat their
water or practice safe water storage and handling. We, therefore, recommend all rural
households to develop the habits of household water treatment to ensure safe water quality—
particularly for drinking. Fifth, households should be made aware of the importance of safe WASH
through educational campaigns so as to help them change their long-held habits and hygiene
behaviors. As the study revealed, when primary caretakers consistently practice handwashing
with soap at critical times and safe child stool disposal, the risk of young children contracting
diarrhea was reduced. Therefore, educating rural communities on the potential sources of water
contamination, proper water treatment methods, safe disposal of feces away from the domestic
environment, and good hygiene practices, such as handwashing with soap at critical times, could
result in significant health gains to the rural population. At the same time, containers used for
collecting or storing water need to be cleaned regularly to ensure safe water quality.
On a policy level, our findings indicate that WASH interventions are also needed to improve
household water treatment, safe sanitation and hygiene practices in rural areas. Primary
caretakers often undermine the critical role that good hygiene plays in improving overall health
outcomes (Roushdy et al. 2012). In addition to affecting health, inadequate water supply may
cause households to limit their handwashing practices and to wash dishes and clothes using
water from unimproved sources. The study also highlights that proper child stool disposal
behavior is lacking. Most of the households perceived child feces to be less harmful than adult
ones and child feces were therefore often left around or disposed close to a household’s living
area. This further highlights the lack of awareness among the study households about the causes
of diarrhea and the necessary remedial measures. Moreover, many primary caretakers do not
consider diarrhea to be a serious health problem as it is common among young children.
Education and public awareness campaigns could be an effective channel to disseminate
information that can reduce child morbidity associated with insufficient WASH. This is particularly
important in many rural areas of the country, where mothers usually have little education. Such
campaigns can be implemented on the ground by health extension workers. Proper hygiene and
childcare practices can be promoted through the Ethiopian government’s Health Extension
Program (HEP) in many rural villages of Ethiopia. A report on rural water supply in Ethiopia found
that most rural water sources were poorly maintained and often contain water unsafe for
drinking; in some instance, water sources were not functioning due to repair and maintenance
75
problems (UNDP 2006). Poor people are the ones suffering the most from the burden of diseases
associated with a lack of improved WATSAN services. The poor also lack awareness about the
detrimental health impacts of poor water quality, unsafe sanitation habits, and inadequate
hygiene practices. Our descriptive statistics provided clear evidence that most rural households
practice unsafe child stool disposal, inadequate household water treatment, and improper
hygiene.
Finally, to address the problem of WATSAN, the GoE has committed by adopting the Universal
Access Plan to achieve 100 percent drinking water coverage at the national level. However, the
overall progress has been slow and there is a disparity in development between urban and rural
areas. The WHO/UNICEF (2015) report on the progress on WATSAN showed that in 2015, 57
percent of the population have access to improved water sources (compared to 13% in 1990) and
28 percent of the population have access to improved sanitation services (compared to 3% in
1990). It is clear that sanitation coverage is lagging far behind water supply coverage. There
should be more concerted and coordinated actions to meet the SDG-6 which aims to ensure
access to WATSAN for all. Unless efforts to increase access to improved WATSAN services are
intensified and implemented in conjunction with the promotion of proper hygiene practices,
communicable diseases will continue to remain a major cause of child morbidity and mortality in
rural Ethiopia.
4.7 A Cost-Benefit Analysis (CBA) for Rural Water Supply in Ethiopia
This section analyzes the cost-effectiveness of investing in rural water supply. We considered a
project to construct a drinking water well fitted with a hand pump in the study areas. Following
Rogers et al. (1998) framework, we assessed the various components of costs and benefits of
water supply as socioeconomic goods.
Welthungerhilfe provided data on capital investments, operation and maintenance (O&M) costs,
the number of beneficiaries, and the lifespan of the well. The followings are assumptions for the
cost-benefit ratio of investment on drinking water well with a hand pump in rural Ethiopia
Average exchange rate $1 = 22.3 ETB (December 2016)
Service life of the well is 10 years
Capital charges (up-front initial investment) = $2,909
Number of beneficiary households per well = 80 (400 people)
Estimated shadow wage for adult woman and man is equal to $0.9, $1.9, respectively.
That is, average shadow wage rate is equal to $1.4
Discount rates 3, 5 and 10 percent
Shadow wage is derived from the marginal productivity of labor based on the estimated
agricultural production function (see column 2 Table A10 in Appendix A), and computed as:
76
𝑀𝑃𝐿𝑖 = �̂�𝑖�̂�
𝐿𝑖, where �̂�𝑖 the coefficient on log (Labori) and Y is the predicted values of output.
a) Costs Analysis
As a rule of thumb, 10 percent of capital charges are considered to be O&M expenses (Cameron
et al. 2011). In this case, however, only the cost of the pump and salary for a security guard should
be used for this calculation, which is much lower than the 10 percent of capital charges.
Considering this, the O&M cost is estimated to be $141 per year. As the O&M cost is a recurring
expense, it needs to be projected over the next 10 years, and the resulting O&M cost is $1,241
using 3 percent discount rate.
The next component of this analysis is the opportunity cost, which is the next best use of an input.
The opportunity cost of an input is the competitive market price of that input. Beneficiaries of a
well project are likely to participate in training for several hours (e.g. health and hygiene
education, technical training). We assume that one working day per household is required at the
time of well construction, and this should be accounted for as an opportunity cost for the
beneficiaries. In this analysis, we used the average shadow wage rate to monetize all time-based
measurements due to lack of competitive market price, and the total present discounted value
(PDV) for the opportunity cost is estimated at $112.
As drinking water contamination both at the POS and POU is high, increasing access to safe water
supply alone is not sufficient to improve health outcomes. We take into account the expenses
required for treatment. Assuming the per capita per day water consumption is treated (8 liters
per day), each person will need 2, 920 liters treated per year. Water treatment cost is estimated
to be $0.70 per cubic meter (Rogers et al. 1998). The benefits for societal objectives can be
significantly low unless water treatment is provided.
The last section of the cost analysis is to determine negative externalities (economic and
environmental costs) associated with the project. As the negative economic externalities for
constructing a well is unclear (Whinnery 2012) and data is not available, a value of $0 is used for
this analysis. On the other hand, although the values are relatively small, there are possibly
significant environmental externalities associated with the construction of a well. These costs are
related to the raw materials and estimated to be $48 per well in east Africa (for a detail
discussion, see Whinnery 2012).
b) Benefits Analysis
Improving access to safe drinking water can generate a broad range of economic and social
benefits. One of the primary benefits of the project is the value to water users. In addition, other
potential net benefits including benefits from indirect uses, benefits from return flows,
adjustment for societal objectives, and intrinsic value can result from improved water supply.
Benefits from return flows, intrinsic value, and indirect use from drinking water supply are set to
77
values of zero because they are often assumed to be insignificant (Whinnery 2012). These
benefits, however, could be much higher in irrigation and hydropower developments projects. In
most cases, it is possible that livestock and/or small-scale agriculture could benefit from an
improved water supply in rural areas. Excluding this and other possible indirect benefits make
the analysis more conservative. This analysis focuses on the direct health impacts of clean water
in terms of associated reduction in diarrhea and related diseases. Although there are other health
issues related to lack of access to clean water, diarrhea is the most significant. This study does
not take into account all secondary benefits or positive externalities to avoid possible double
counting. Other direct benefits, such as money saved on medical expenses, are not included in
this section either. Generally, such an approach only makes this analysis more conservative.
For estimating the value to water users, this analysis uses monetary value of time saved due to
improved access to clean water supply for assessing the value of the project to its beneficiaries.
In this case, data is available that estimates the average number of water collection trips per
week and the time needed for a round water collection trip in rural households. An average
household spends 2.36 hours per day for water collection activities. Multiplying this figure by 365
and then by the number of total beneficiary households gives the total hours spend for water
collection. Furthermore, to monetize the value of time-saved, we use the average shadow wage
rate ((1/8*2.36*80*365)*$1.4 = $12, 060. This benefit needs to be discounted for annual benefits
in the future, and the resulting PDV of time saved is $114,934. The benefits of time saved due to
a new community water source highly depend on the population density surrounding the
location of the well. As populations are sparsely settled in rural areas, assuming zero water
collection time for all the beneficiary households would over-estimate the value of users to
water. However, this could be partly compensated by their willingness to pay the available water
services, which we did not take into account.
Another important benefit from this project is to improve the health of community currently
lacking access to safe drinking water; this might be the central reason for many organizations
working in this sector. These key societal objectives can be measured using the concept of value-
of-statistical-life (VSL), which VSL places an economic value to life lost/saved and/or productivity
gain/reduction for a given individual. Viscusi and Aldy (2002) estimated the VSL in the United
States at $8.9 million. To value the health impacts of the project (the monetary values estimates)
the VSL must be set. The VSL in Ethiopia can be estimated by making appropriate adjustments
(using GDP or GNI per capita), between the US and Ethiopia, and this results in a VSL of $257,018.
The next step is to estimate the value of a disability-adjusted life year (DALY). The DALY approach
is designed to quantify the impact of various diseases on the productivity and well-being of
people by estimating the monetary values (Li 2014). It is, however, highly controversial and
criticized by many in the evaluation field (for a detail discussion, see, Li 2014). It has been
criticized for discriminating the disabled people, that is, the lives of disabled people are less
78
valuable than the lives of people without disability. This unequal valuing of life is unethical and
discriminatory (Bickenbach 2008). Despite demonstrated flaws in its intrinsic assumptions and
justification, the DALY is still widely used in the field of public health to estimate the cost-
effectiveness of different programs. Disability-adjusted life year (DALY) combines all associated
health costs, including Years Lost due to Disability (YLD) for people living with the diseases and
Years of Life Lost (YLL) due to premature mortality. The World Health Organization (2012)
commonly uses an average equivalency of 36 DALYs per premature death. Again replicating the
same approach, the value per DALY can be determined by dividing the VSL by the number of
DALYs assigned for a premature death, and this results in the value of $7,139 per DALY in Ethiopia;
see Table 4-10.
Table 4-10: Metrics summary of value per DALY calculation
GNI per capita, PPP adjusted
Value of a Statistical Life (VSL)
DALY equivalent for a premature death, on average
Value per DALY
US $56,189 $8,914,553 36 $247,626
Ethiopia 1,620 $257,018 36 $7,139
Source: Column 1 (World Bank 2015); column 2 (Viscusi & Aldy 2002); column 3 (WHO 2012).
To quantify the number of DALYs averted due to this project, we have to determine the number
of YLD and YLL averted. In our sample, the average number of diarrhea cases per child per year
is 4.8, and a case of diarrhea lasts 3.4 days. The disability weight for diarrhea is 10 percent
(Cameron et al., 2011; Pruss et al. 2002). The number of beneficiary households is assumed to be
80 for a well. The average number of under-five children is 1.24, that is, about 100 children can
benefit from the project. The number of morbidity-based DALYs averted by this program is
estimated to be 0.067 per year.
YLD = (100 beneficiaries) (4.8 diarrhea cases per year) (3.4 days per case) (1/365) (10% disability
weight) (15% reduction)
As reported by the WHO, an average of 64 lives per 1000 under-five children are lost each year
in Ethiopia, of which nine of them is due to diarrheal diseases alone. This is the national average,
however, the number of children lost due to diarrhea could be much higher in rural areas
compared to urban. Assuming improving access to drinking water reduces deaths caused by
diarrhea at the same rate as the incidence reduction (15 percent), 0.135 mortality-based DALYs
are averted each year.
YLL = (100 beneficiaries) (9/1000 lives lost) (15% reduction)
In addition to these averted deaths, there could be additional averted deaths due to access to
safe drinking water through indirect/secondary pathways. Adding up the two components, 0.202
DALYs are averted annually that can be attributed to improved access to drinking water.
79
Discounting for future benefits over the anticipated life of the project, the PDV of the adjustment
for societal objectives is $13,749. For the computation see Table 4-11.
Moreover, health benefits for non-preschool household members and all indirect health impacts
including health externalities of communicable diseases were not fully accounted for. Hence if all
the possible benefits could be monetized, the benefit-cost ratios would be considerably higher
than the one presented. Finally, despite our conservative assumptions, this study has confirmed
that drinking water supply in rural areas generates high economic returns, with returns far
exceeding costs even at 10 percent discount rate. For instance, the results suggest that for every
US$1 invested in rural water supply, there is an economic return of US$19 by saving water
collection time and keeping people healthy and productive. The estimated cost-benefit ratio is
relatively higher than what is being reported by Hutton and Haller (2004). They estimated cost-
benefit ratio between US$10 and US$15 for improved water supply and sanitation in developing
countries. While the aforementioned study is at a global level based on the WHO sub-regions,
this is a case study in rural areas where access to drinking water supply and sanitation is low and
diarrhea prevalence is high. In such circumstances, the time benefits from the proximity of water
supply and the health benefits from reduced morbidity and premature child mortality due to
diarrheal diseases is expected to account for a large share of expected total benefits (Hutton &
Haller 2004).
Table 4-11: Metrics and summary of value per DALY calculation
YLD (Morbidity) YLL (Mortality)
Average annual diarrhea cases per child
4.8 Proportion of lives lost to diarrhea each year per 1000 child
0.009
Days lost per case of diarrhea 3.4 Reduction in diarrhea due to intervention (Table 4.7)
15% Disability weight for diarrhea 10%
DALYs averted each year 0.067 DALYs averted each year 0.135
Value of YLD averted each year $478 Value of YLL averted each year $964
Source: Author’s own computation.
80
Table 4-12: Summary of cost-benefit analysis
Costs Descriptions Estimated values Reference(s)
Capital Charges well construction & development $2,909 Welthungerhilfe
O&M Cost PDV assumes 5% capital cost as proxy
for annual O&M cost, including security guard
$1,241 Welthungerhilfe
Opportunity Cost time spent for training (1 day per beneficiary)
$112 estimated based
on the market assumption
Economic Externalities negative economic impacts $0 n/a
Environmental Externalities negative environmental impacts $48 see Whinnery 2012
Water treatment assume 8 liters per user per day need
to be treated, and the cost is $0.70 per cubic meter (1000 liters)
$1,944 Reed 2005;
Roger et al. 1998
Total PDV Costs $6,254
Benefits Descriptions Estimated values Reference(s)
Value to Users of Water PDV of Value to Users of Water $114,934 estimated
Net Benefits from Return Flows
n/a $0 n/a
Net Benefits from Indirect Uses
n/a $0 n/a
Adjustment for Societal Objectives
PDV of DALYs Averted $13,743 estimated
Intrinsic Value n/a $0 n/a
Total PDV Benefits $128,677
NPV (total PDV Benefits - total PDV Costs) $122,423
Benefit/Cost Ratio (discount rate 3%) 20.58
Benefit/Cost Ratio (discount rate 5%) 19.72
Benefit/Cost Ratio (discount rate 10%) 17.80
Source: Author’s own computation.
81
Table 4-13: Present discounted value (PDV) computation details
Year PDV
O&M PDV Water Treatment PDV Value to Users of Water
PDV DALYs Averted
0 n/a $204.00 $12,067.00 $1,442.64
1 $141.21 $198.06 $11,715.53 $1,400.63
2 $137.10 $192.29 $11,374.30 $1,359.83
3 $133.11 $186.69 $11,043.01 $1,320.22
4 $129.23 $181.25 $10,721.37 $1,281.77
5 $125.47 $175.97 $10,409.10 $1,244.44
6 $121.81 $170.85 $10,105.92 $1,208.19
7 $118.26 $165.87 $9,811.58 $1,173.00
8 $114.82 $161.04 $9,525.80 $1,138.84
9 $111.48 $156.35 $9,248.35 $1,105.67
10 $108.23 $151.80 $8,978.98 $1,073.46
Note: Discount rate = 3% and 𝑃𝐷𝑉 = ∑𝐷𝑡
(1+𝑟)𝑡𝑛𝑡=0
82
Appendix 4
Figure A1: Proportion of children with diarrhea by number of episodes and irrigation
Figure A2: Proportion of children with diarrhea by episode duration and irrigation
Source: Author’s computation using own survey data
0
0.1
0.2
0.3
0.4
0 1 2 3 4 5 6
Pro
po
rtio
n o
f ch
ildre
n w
ith
dia
rrh
ea
Number of episodes
Irrigator Non-irrigator
0.00
0.10
0.20
0.30
0.40
1 2 3 4 5 6 7 8 9 >=10
Pro
po
rtio
n o
f ch
ildre
n w
ith
dia
rrh
ea
Episodes duration (days)
Irrigator Non-irrigator
83
5. Water, sanitation, and agriculture: Linkages and impacts on health outcomes
5.1 Introduction
The interactions between irrigated agriculture and water and sanitation (AG-WATSAN) are
complex where water is an important determinant factor for livelihoods. As water is a key aspect
for subsistence farmers, the implications of lack of water access are extremely broad. In a country
where agriculture is the backbone of the economy, water shortages mainly affect not only its
economy but also the lives of many smallholder farmers whose livelihoods entirely depend on
crop production. Consequently, supporting small-scale irrigation (SSI) has been identified by the
GoE as a key poverty reduction strategy to overcome existing challenges and to build household
resilience and sustainable solutions.
The country’s Agricultural Growth Program (AGP) is focused on improving water availability for
small-scale agriculture in rural areas where people’s livelihoods highly depend on water. Water
resource management is, therefore, vital for transforming rain-fed farming to irrigated
agriculture thereby improving productivity in a sustainable manner. Irrigated agriculture is
expected to increase agricultural production and productivity through the use of water resources
such as surface, ground and rainwater and use of appropriate input mix (Asayehegn et al. 2012;
Burney et al. 2010; FAO 2000). In the last few years, more and more small-scale farmers have
been adopting irrigation technology to increase production on top of subsistence farming.
Improved rural water supply coverage is also inadequate in the country: if any, it is basic and
inadequate (van Koppen, Smits, Moriarty, et al. 2009; Usman et al. 2016). Consequently, separate
sources of water for household and water for agricultural uses do not exist in several rural areas.
On the one hand, domestic water supply is often used for watering livestock, SSI and gardening,
and other income generating activities (Butterworth et al. 2013; Scheelbeek 2005; van Koppen,
Smits, & Mikhail 2009). On the other hand, as access to safe drinking water supply is inadequate,
households may be often forced to use irrigation water for a wide range of domestic purposes.
There is, however, a trade-off in the domestic use of irrigation water. On the one hand, irrigation
water might serve as an option to increase the availability of water for domestic uses and saves
water collection time and energy. On the other hand, use of poor quality irrigation water for
domestic purposes is harmful to health due to the presence of disease-causing pathogens.
Nevertheless, the empirical evidence on the health impacts of multiple-use of irrigation water is
scant (Boelee et al. 2007; Tsegai et al. 2013). Furthermore, Boelee et al. (2007, p.49) recommend
that “the interactions between irrigation and domestic water supply, water quantities involved,
hygienic behavior, and mechanisms of disease transmission are poorly understood.” They
suggest that more field studies are needed particularly in sub-Saharan Africa to identify the right
technical and institutional arrangements under which irrigation and domestic water supply could
84
improve human health and rural livelihoods. This study, therefore, attempts to fill this research
gap by examining the linkages and interactions between water supply for domestic use and for
agricultural production, and their health and nutrition impacts on households under a multiple-
use water system setting.
5.2 Linkage between WASH, Health, and Agriculture
In rural communities, irrigated agriculture has a complex interaction with drinking water supply
and sanitation services. Use of irrigation water for the domestic purpose can have some negative
effects but it also has positive impacts in terms of improving human health and nutritional status.
Increasing water availability—a critical linkage to hygiene—very close to the house primarily
allows a given household to collect enough water for better hygiene practices in addition to
overcoming a lack of basic needs. This prevents diseases that are related to the amount of water
available for personal hygiene. In addition to the health benefits of a large amount of water
availability for domestic uses, there may be other indirect effects by reducing water fetching time
and energy, which in turn increases time availability for, for example, childcare, food preparation
and engaging in other productive activities (Dufour et al. 2012; Hagos et al. 2008; Koolwal & Walle
2013; Pickering & Davis 2012; Wang & Hunter 2010).
Irrigation impacts human health in several pathways. Irrigated agriculture often increases
productivity (production) and creates opportunities for diversification of livelihoods that allows
households to have better nutrition and earn a higher income (von Braun et al. 1989). The use of
water resources for productive activities has high value in terms of household income, nutrition,
and health. In addition to providing more nutrient-rich foods and vegetables for household
consumption, irrigation activities generate higher income that allows access to improved health-
care services. Therefore, irrigation has an indirect effect on health via income. However,
irrigation systems might have other downside effects and can be sources of health hazards-
especially in areas prone to waterborne diseases as standing water provides a conducive
environment for diseases (for example, malaria). In Ethiopia, several empirical studies revealed
that the prevalence of malaria and water-related disease are higher in villages closer to micro
dams or irrigation systems (Amacher et al. 2004; Asayehegn 2012; Ersado 2005), which is among
the top five deadly diseases in the country. However, a study in Tanzania shows that malaria
prevalence in irrigated areas was lower because the improved income from rice-growing enables
the farm-households to afford to use insecticide-treated nets (Mutero et al. 2006). Malaria and
other waterborne diseases affect the person’s ability to work and reduce the time available for
productive activities. Other members of the household will have less time for productive
activities as they have to spend time away from work to take care of ill relatives. This affects the
income of the household directly due to an increase in medical expenses and indirectly by
reducing their ability to work on their farm yards.
85
It is estimated that more than 52 percent of rural population in sub-Saharan Africa alone could
benefit from the multiple-use water system if rural water supply is designed to accommodate
additional services beyond the domestic water supply (Faurès et al. 2008). In summary, most
existing empirical studies suggest that water resources and irrigation systems management
should be given high importance to avoid the unpremeditated negative health repercussion for
the most vulnerable people and communities while exploiting the opportunities opened up to
change and improve the livelihood of the poor at large.
5.3 Theoretical Framework
This analysis draws upon the household production model (Becker 1965) in which households
are assumed to maximize their utility function subject to a budget and a time constraint. In line
with the theory, households purchase different goods and combine them into a household
production system to produce various goods and services under a given time constraint. These
purchased goods and produced commodities directly enter into the household’s utility function.
Since most human capital outcomes are not available in the market, households have to produce
them on their own based on the integration of biological, demographic, and economic
considerations.
Consider a one-period household production model with constrained maximization of a joint
utility function,31 that is, non-separable determination of household production—consumption
decision. Extending the exposition set out in Behrman and Deolalikar (1988), assume that the
household preferences constituting of T individuals are represented by the following preference
function:
U = U(𝑋𝑖, 𝐶𝑖 , 𝑇𝑙𝑖 , 𝐻𝑖)𝑖 = 1,… , 𝑇 (5.1)
where 𝑋𝑖 is the consumption of purchased goods of household member i,𝐶𝑖 is the consumption
of own production of household member i,𝑇𝑙𝑖 is the leisure time of household member i,𝐻𝑖 is
the health of household member i. The utility is assumed to depend on the consumption of goods
(both markets purchased goods and own produced agricultural goods), the leisure time and the
health status of each of the household members.
The preference function in equation (5.1) is maximized subject to the following constraints. The
first constraint is a health production function. The health of the ith individual is produced by the
consumption of goods and leisure. Health also depends on the consumption of non-food health
inputs which do not provide utility directly (such as medical treatment from local clinics); a vector
of household resources (such as drinking water and sanitation facilities); a vector of irrigation
31 The assumption of “unitary” household preference approach may be questioned, but it is still unclear how a
cooperative bargaining process affects our outcome variable and changes the theoretical pathways through which improved WATSAN affects health (for further discussion see Alderman et al. 1995; Mwabu 2007).
86
water characteristics (such as types of irrigation water, distance from the household); the i’s
innate health endowment and specific individual characteristics, and a vector of community
characteristics that include variables that affect an individual’s health (such as access to health
facilities, infrastructure and prices, cleanness of the environment).
Irrigation technology can improve health and nutrition by providing nutrient rich and fresh
vegetables. It can also improve health by providing greater water availability for domestic use
and enables households to maintain key hygiene and sanitation practices (van der Hoek et al.
2002). However, irrigation intervention can affect health negatively in areas prone to vector-
borne diseases (Amacher et al. 2004; Ersado 2005).
The health production function then is given by:
𝐻𝑖 = H(𝑋𝑖, 𝐶𝑖 , 𝑍𝑖 , 𝑇𝑙𝑖 , 𝑇ℎ
𝑖 , 𝐺, 𝐷, 𝜂𝑖, Ω) (5.2)
where 𝑍𝑖 is the consumption of non-food health inputs of household member i (e.g. health care
services), 𝑇ℎ𝑖 the time spent on production of health (e.g. taking care of ill family members), 𝐺 a
vector of household resources which affect health (e.g. water & sanitation facilities), 𝐷 a vector
of irrigation water characteristics (e.g. its type and distance from the dwelling unit where the
individual lives), 𝜂𝑖 the observed and unobserved individual characteristics, and Ω the
endowment of the households and community characteristics (e.g. the general environment),
and the other variables are defined above.
The production/consumption of water for domestic use depends upon the quality of the water
source, the time spent by the ith household member collecting water and is assumed to be a
function of distance to the water source; the knowledge of good health practices in the
household as they relate to water collection and storage, the capital goods used in the transport
and storage of water.32 Water quality at the source and quality of water at the POU may
substantially vary. The average water quality at the POU can be lower relative to the quality of
water at the source as it is affected by factors such as transportation, knowledge of safe water
handling, the quality of storage containers and duration time (for a discussion see Wright et al.
2004; Zwane & Kremer 2007).
The second constraint is a water production function that can be defined as follows:
𝑊ℎ = W(ϖ, 𝑇𝑤𝑖 (𝐷𝑤), 𝐴𝑤, 𝐻
𝑖, 𝐷, Ω) (5.3)
where ϖ is the quality of water available to the household member i, 𝑇𝑤𝑖 the time spend by
household member i on water collection, 𝐷𝑤 the distance to water source, 𝐴𝑤 the capital goods
32 It is possible that and Dw become a choice variable if households can choose among different sources of water
for domestic consumption.
87
used in the transport and storage of water (e.g. buckets and pots), and the other variables are
defined above.
The household farm production depends not only on the characteristics of all individuals in the
household who work on these activities but also on the amount of land, the levels of input use
and the practice of irrigation technology. A study by von Braun et al. (1989) shows that utilization
of irrigation technology in the production of rice in Gambia increases not only the household’s
real income but also it increases their calorie consumption and food expenditure. Agricultural
production can also be affected by the household members’ health status either through its
effect on the quality of own labor or through reducing labor availability.
The third constraint is a farm production function:
𝑌ℎ = Y(𝐶𝑖, 𝑇𝑓𝑖 , 𝐿ℎ, 𝐴, 𝐾, 𝐻
𝑖, 𝐷, Ω) (5.4)
where 𝑌ℎ is the household farm output aggregated overall crops and vegetables, 𝑇𝑓𝑖 the time of
the ith household member spent on household farm production, 𝐿ℎ the amount of hired labor
used in farm production, 𝐴 the amount of land used in the farm production process, part of which
may be owned by the households (Ah) and the rest of which may be rented (A*), 𝐾 other variable
inputs used in farm production and all other variables have been defined above.
Finally, there are time and full-income constraints. Household’s total labor time (T) available is
allocated to leisure, health-care activities, water collection, and agricultural production and off-
farm income generation:
T = 𝑇𝑙𝑖 + 𝑇ℎ
𝑖 + 𝑇𝑤𝑖 + 𝑇𝑓
𝑖 + 𝑇𝑜𝑓𝑓𝑖 )∀𝑖 (5.5)
Continuing with the description of the model, it is also assumed that households face a monetary
budget constraint where income equals expenditure for consumption goods, leisure, health-care
services, expenditure for water container, hired labor is represented by a full-time income
constraint as:
𝑃𝑦𝑌ℎ + R = rA + 𝑃𝑘𝐾 + 𝜔𝐿ℎ + 𝑃𝑥𝑋 + 𝑃𝑐𝐶 + 𝑃𝑧𝑍 + 𝑃𝑤𝑊
+∑𝜔(𝑇𝑖 − 𝑇𝑙𝑖 − 𝑇ℎ
𝑖 − 𝑇𝑤𝑖 − 𝑇𝑓
𝑖 − 𝑇𝑜𝑓𝑓𝑖 )
(5.6)
where r is the rental rate of land, R other exogenous income, 𝜔 the market wage rate, 𝑇𝑖 total
time of the ith individual, 𝑇𝑜𝑓𝑓𝑖 the labor market work time of the ith individual, 𝑃𝑗 refers to the
different prices, where j = y, k, x, c, z and w.
Maximizing equation (5.1) subject to the (5.2), (5.3), (5.4) and (5.6) constraints yields the
following reduced-form demand function for all choice variables in which all exogenous variables
88
appear on the right-hand side of each equation, and the left-hand-side variables are the
endogenous variables in the system for the household. Therefore, the reduced-form demand
functions for our choice variables are given as:
𝑉 = 𝑓(𝐸) (5.7)
where 𝑉 = (𝐻𝑖,𝑊ℎ , 𝑍𝑖 , 𝑌ℎ, 𝑋𝑖 , 𝐶𝑖 , 𝑇𝑗𝑖)𝑤ℎ𝑒𝑟𝑒𝑗 = 𝑙, ℎ, 𝑤, 𝑓, 𝑜𝑓𝑓; and
𝐸 = (𝑟, 𝑃𝑠 , 𝜔, 𝐺, 𝐷𝑤, 𝑅, 𝐷, 𝜂𝑖 , Ω)𝑤ℎ𝑒𝑟𝑒𝑠 = 𝑎, 𝑐, 𝑤, 𝑥, 𝑦, 𝑧
This theoretical framework tries to show the nexus among WATSAN, irrigation and health
functions. This framework helps us to identify some policy variables that affect the demand for
health, water, sanitation and agricultural irrigation system. Therefore, estimation of the reduced-
form demand functions provides a consistent framework.
5.4 Data and Empirical Strategy
5.4.1 Data
The data come from the same household survey used in the previous empirical chapters. The
descriptive statistics for the combined sample has been presented in the previous chapters (see,
Table 4-3 and Table 4-3). Table 5-1, however, presents the summary statistics of the variables
of interest by irrigation status.
The descriptive statistics presented in Table 5-1 shows that the prevalence of diseases appears
significantly higher among irrigator households than non-irrigator households in all health
outcomes measures except for diarrhea. Moreover, health expenditure is higher among irrigator
than non-irrigator households, and this difference is statistically significant at the 1 percent
significance level. Regarding access to improved WATSAN, the result shows that irrigator
households had better access to improved water source than non-irrigator households; however,
storage water quality is slightly better for non-irrigator than irrigator households. Furthermore,
the l/p/c/d water consumption is relatively higher among irrigator households which may be
because irrigator households spent lower time per water collection trip than non-irrigators. We
also observed difference between irrigator and non-irrigator households in terms of access to pit
latrine, and this significant difference in terms of access to pit latrine can be interpreted as higher
proportions of irrigator households practicing open defecation than non-irrigator households. On
the other hand, as expected, household income from agriculture (excluding livestock) is much
higher among irrigator than non-irrigator households. The table also shows that 66 percent of
the total sampled households engaged in irrigated agriculture.
89
Table 5-1: Descriptive statistics
Variables Total
sample N Irrigator household Non-irrigator household
mean Mean SD Mean SD
Health outcomes
Diarrhea prevalence 0.09 2149 0.10 0.30 0.09 0.28
Malaria incidence 0.04** 2149 0.05 0.21 0.03 0.17
Water-related diseases a 0.11** 2149 0.13 0.33 0.08 0.28
Overall morbidity b 0.31*** 2149 0.34 0.47 0.26 0.44
Total medical expenditures in the previous two months (ETH BIRR)
84.36*** 454 102.83 201.93 47.64 107.87
Healthcare expenditures in the last one year (ETH BIRR)
268.89*** 454 313.51 421.59 180.24 214.05
Water supply and sanitation
Access to improved water source 49.56 454 0.51 0.50 0.46 0.50
Storage water quality (1=no E.coli) 0.42 454 0.41 0.49 0.43 0.49
Level of E.coli (CFU/100ml) water 23.31 454 24.72 40.75 20.52 34.16
Minutes to water sources 20.38*** 454 17.18 13.56 27.10 17.48
Per capita per day water consumption in litters (l/p/c/d)
8.77*** 454 8.96 2.36 8.39 2.36
Soap availability (1= yes) 0.76 454 0.76 0.43 0.77 0.42
Access to pit latrine (1= yes) 0.44*** 454 0.40 0.49 0.52*** 0.50
Socioeconomic and demographic characteristics
Household head age 37.72 454 37.98 0.51 37.19 0.67
Household head’s education 1.02 454 0.10 0.11 1.07 0.16
Highest education completed 3.5* 454 3.37 0.18 3.76 0.24
Household size 5.98* 454 6.08 0.10 5.79 0.14
Number of children under 5 years 1.24 565 1.26 0.47 1.22 0.41
Agri. land size in hectare (operated) 1.31 454 1.29 0.04 1.34 0.06
Agricultural income in 1000 Birr 7.58*** 454 10.17 8.53 2.14 2.22
Non-farm income in 100 Birr 7.57 454 8.25 21.34 6.21 15.08
Community characteristics
Distance to the nearest asphalt road in km 5.96** 454 6.45 0.33 4.98 0.67
Distance to the nearest periodic market km 4.40*** 454 5.05 0.26 3.10 0.38
Altitude above sea level (meters) 1902.56*** 454 1857.00 4.25 1993.07 8.64
a Water-related disease includes malaria, pneumonia, skin itching, typhoid fever, trachoma & giardiasis. b Morbidity includes any diseases symptoms except injuries. Significant difference at *** p<0.01, ** p<0.05, * p<0.1
Source: Author’s estimates using own survey data.
90
Outcome variables
We have proposed an array of outcome variables to measure the health status of a household.
In this section, we use diarrhea prevalence as the main outcome variable because it is strongly
associated with drinking water quality. For robustness test, we also use water-related diseases
(such as malaria, pneumonia, skin-itching, typhoid fever, giardiasis, and trachoma) and morbidity
(general illness of household member excluding injuries) as health indicators for non-preschool
household members. Data on the health status of all household members in the preceding two-
months before the survey were collected.
5.4.2 Empirical strategy
This section outlines the empirical strategy we use to estimate the impact of irrigation water use
for domestic purposes on household health status. Following equation (5.7), the reduced forms
of h*(health) and 𝑊ℎ(household water) are influenced by D (a vector of irrigation
charachterstics), and a set of other explanatory variables. Hence, we are interested in estimating
the effects of domestic use of irrigation water on household health, the reduced form equation
can be expressed as:
ℎ = 𝛽𝜒 + 𝛿𝑄 + 휀 (5.8)
where h denotes health outcomes and X is a vector of exogenous explanatory variables that
influence the outcome variables, and it includes household and individual characteristics, such as
age, gender, household size, among others, and their impact is captured in the vector of
parameters β. Q is a dummy for households directly withdrawing water from irrigation sources
for domestic purpose and the coefficient δ, measures the impact of irrigation water use on
households’ health status, ε is the error term. The variable, Q, is potentially endogenous since
irrigation water for domestic use is not randomly assigned, and households may decide whether
or not to use irrigation water for domestic purposes (i.e. self-selection bias), that is, households
using irrigation water for domestic consumption may be systematically different from the
comparison group households for several reasons. If there is a self-section problem and/or if any
unobserved household characteristics influence both the outcome and the variable of interest—
domestic use of irrigation water, δ will not capture the true impacts of domestic use of irrigation
water on households’ health outcomes. As a result, estimating equation 5.8 without controlling
for such potential endogeneity may yield biased results, and the estimated coefficient hardly be
interpreted as an impact. To account for this, an instrumental variable approach is often
suggested. As we are constrained with a valid instrument to use a two-stage least squares
estimation strategy, we estimated several models with different specifications to test the
robustness of estimated results.
91
In this chapter, we mainly focused on the sub-sample households, who have already adopted
irrigation technology, to investigate the effects of domestic use of irrigation water on health
status and drinking water quality/quantity. However, we also briefly investigated the role of
irrigation on household income, health and domestic water quantity using the whole sampled
households.
5.5 Results and Discussion
5.5.1 The influence of demographic and WATSAN characteristics on diarrhea and other water-
related diseases
The incidence of diarrhea in the last two months prior to the survey among non-preschool
household members by demographic and WASH characteristics is presented in Table 5-2. The
table shows that individual‘s age highly influences the prevalence of diarrhea. Diarrheal incidence
was much higher among young household members—particularly among those young children
between 5 and 6 years old followed by children between 7 and 9 years old. Moreover, female
household members are more likely to be affected by diarrhea than male household members.
In addition, a lower incidence of diarrhea was reported among individuals living in households
where household’s head or primary caretaker is literate. The result also shows that individuals
living in households with improved drinking water source are less likely to have had diarrhea in
the past two months (6.20%) than those living in households with unimproved water sources
(12.55%). The result is same when disaggregated by actual water sources, that is, household
members with unprotected well/spring or surface water sources are more likely to experience
diarrhea than household members with protected private/shared wells. Similarly, higher
diarrhea incidence was reported among households whose POU (storage) water is contaminated
with E.coli than households with uncontaminated POU water (5.36% compared to 12.29%). The
result also shows that soap availability is associated with lower diarrhea incidence. On the other
hand, the difference in the prevalence of diarrhea in the last two months before the survey
between households with and without a pit latrine is negligible. Moreover, diarrhea incidence
was slightly higher among irrigator households than non-irrigator households (9.76% compared
with 8.65%).
92
Table 5-2: Diarrhea incidence by demographic and WASH characteristics (last two months)
Variables Incidence of diarrhea* (%) Number of individuals
Demographic characteristics
Age in years
5 - 6 36.41 217
7-9 15.85 347
10 -15 6.18 453
16 years/older 3.53 1132
Sex
Male 8.82 1134
Female 10.05 1015
Primary caretaker’s literacy
Yes 5.75 174
No 9.73 1974
Household head’s literacy
Yes 7.84 1008
No 10.43 1141
WASH characteristics
Water source based on WHO
Improved source 6.20 1065
Unimproved source 12.55 1084
Water sources
Private protected dug well 6.03 116
Shared protected dug well/spring 6.22 949
Unprotected well/spring 12.62 848
Surface water 12.29 236
Variables Incidence of diarrhea* (%) Number of individuals
Storage water quality
Uncontaminated 5.36 896
Contaminated (at least one E.coli) 12.29 1253
Soap availability
Yes 8.69 1645
No 11.71 504
Pit latrine
Yes 9.64 944
No 9.21 1205
Irrigation farming
Yes 9.76 1455
No 8.65 694
Note: * Only non-preschool household members.
Source: Author’s estimates using own survey data.
93
Table 5-3: The influence of demographic and WASH characteristics on water-related diseases and overall morbidity (last two months)
Variables Incidence of other water-
related diseases (%) Overall Morbidity
(%) Number of individuals
Demographic characteristics
Age
5-6 7.37 50.23 217
7-9 7.49 29.68 347
10-15 10.60 23.40 453
16 years or older 13.69 31.45 1132
Sex
Male 11.33 29.63 1134
Female 11.46 33.30 1015
WASH characteristics
Water source based on WHO
Improved source 11.55 29.11 1065
Unimproved source 11.25 33.58 1084
Water sources
Private protected dug well 16.38 31.90 116
Shared protected dug well/spring 10.96 28.77 949
Unprotected well/spring 10.14 32.78 848
Surface water 15.25 36.44 236
Storage water quality
Uncontaminated 10.04 27.12 1253
Contaminated 12.37 34.40 896
Irrigation farming
Yes 7.15 34.02 1455
No 4.61 25.79 694
Note: Water-related diseases include malaria, pneumonia, skin itching, typhoid, trachoma, and giardiasis for non-preschool household members.
Source: Author’s estimates using own survey data.
The incidence of other water-related diseases (such as malaria, pneumonia, skin-itching, typhoid,
trachoma, and giardiasis) and overall morbidity (excluding injuries) in the last two months prior
to the survey by demographic and WASH characteristics is presented in Table 5-3. As we can see,
younger household members are the least affected groups in terms of incidence of water-related
diseases; however, looking at the overall health status in the last two months, they are the most
affected segments of the population in the study areas. Male and female household members
were equally affected by water-related illness; nevertheless, male household members are
relatively healthier than female household members in the previous two months. Moreover,
94
household members with improved water sources are slightly healthier than households without
improved water. Surprisingly, however, the result presented in in Table 5-3 indicates that
individuals living in households using a protected private-well or surface water sources are more
likely to have had health problems in the past two months than those living in households using
either a shared-protected or unprotected-well. On the other hand, as expected, the incidence of
water-related diseases and morbidity status are significantly higher among households having
contaminated storage water and engaging in irrigation farming.
The following Figure 5-1 presents the incidence of diseases among non-preschool household
members in the preceding two months before the survey by income quintile (we used adult
equivalent total household expenditure to construct the income quintile). Although
observed household income variability is high, the result shows that income does not play a
significant role in the determination of household members’ health status. That is, both
poorer and richer households are fairly equally prone to diarrhea and other health problems.
In other words, higher income does not make household members immune from contracting
diarrheal and other infectious diseases in the study areas. The role of income on household
health status may also be moderated by poor access and utilization of health facility in the
study areas. It is argued that increased household income permits a given household to
access improved health services and improve their health status. However, this might not
hold where access and quality of health facilities are under developed. In addition, there
might be measurement issues related to household income.
Figure 5-1: Prevalence of illness in the last two months by income quintile
Source: Author’s estimates using own survey data.
8.33 10.98 9.89 8.33 9.48
12.27 8.88 10.11 13.43 12.32
33.128.04
32.64 31.94 31.04
0
5
10
15
20
25
30
35
40
45
50
55
1st 2nd 3rd 4th 5th
Shar
e o
f si
ck h
ou
seh
old
mem
ber
s
Income quintile
Diarrhea Water related Mobidity
95
5.5.1.1 The interaction between irrigation and domestic water supply
Small-scale irrigation practices
In the study areas, farmers practice similar irrigation methods. Furrow and surface irrigation
methods are the dominant irrigation methods practiced by most farmers. Furrow irrigation
method is a “partial surface flooding method of irrigation normally used with clean-tilled crops
where water is applied in furrows or rows of sufficient capacity to contain the designed irrigation
system” whereas surface irrigation method is “where the soil surface is used as a conduit, as in-
furrow and border irrigation as opposed to sprinkler irrigation or sub-irrigation.” 33 In our sample,
59 percent of households use furrow methods, 23 percent of households practice surface (run-
of-river) irrigation type, and the remaining 18 percent of households use manual-bucket
irrigation (Table 5-4). Surface and Furrow irrigation methods are almost similar, while manual-
bucket is a drip irrigation method and mostly used to irrigate small plot of lands.
Table 5-4: Types of irrigation mostly practiced by farmers
Irrigation type Frequency Percent Cumulative
Surface –run of river 77 23.48 23.48
Furrow system 193 58.84 82.32
Manual-buckets /drip 58 17.68 100.00
Total 328 100.00
Source: Author’s estimates using own survey data.
Households use various water sources to irrigate their farm. Rivers/streams, canal water, ground
water, natural pond and harvested water used as a source for irrigation farming (Figure 5-2). For
instance, rivers are the main irrigation water sources in Fogera district, whereas canal water is
the principal water sources for most farmers in Mecha district—Koga irrigation dam. However,
concurrent use of several water resources for different purposes and/or during a different time
of the years is common in these areas. Moreover, WATSAN and irrigation are interlinked as water
from rivers, springs and groundwater are used as a source for irrigation as well as for domestic
consumption and on-site sanitation, which is common among rural households.
33 http://water.usgs.gov/edu/irquicklook.html (accessed on February 2016).
96
Figure 5-2: Primary water sources for irrigation purpose
Source: Author’s computation using own survey data.
Smallholder farmers use a single or a combination of water-lifting technologies to irrigate their
farm plots (Table 5-5). River and stream diversions using gravity are the most widely used
practices to deliver water from sources to the farm fields. They are technically simple and
reliable. However, some farmers also use motor pumps where diversion by gravity is impossible.
Motor pumps can be an appropriate water-lifting technology for groundwater utilization but it
may not be economically affordable for most smallholder farmers. Moreover, buckets are also
used to withdraw water from groundwater and to irrigate homestead plots. Inappropriate water
withdrawal technologies for multiple-use water services can also increase water source
contamination.
Table 5-5: Main types of water delivery system from irrigation water source
Water delivery system Frequency Percent Cumulative
Motor pumps using diesel 109 32.54 32.54
Diversion using gravity 168 50.15 82.69
Manual-buckets 58 17.31 100.00
Total 335 100.00
Source: Author’s estimates using own survey data.
134
82
66
40
70
20
40
60
80
100
120
140
River/stream Dam/canal Groundwater/well
lake ornatural pond
Harvestedwater
Nu
mb
er o
f h
ou
seh
old
s
97
As Figure 5-3 shows, more than 50 percent of households reported that their primary source of
drinking water on the field is carried from home. However, above 24 percent of households
reported that they do not have a separate source of drinking water other than their irrigation
water. The remaining households obtained their drinking water from a nearby protected or
unprotected wells/springs.
Figure 5-3: Primary source of drinking water on irrigation site
Source: Author’s computation using own survey data.
Figure 5-4 shows the major problems related to irrigation water sources in the study areas. Most
households reported that they experience water shortage (lack of access to year-round water
sources). Some households acknowledged that their irrigation water is of poor quality, while
others reported that irrigation water sources create infectious diseases or intensifying the
breeding of mosquitoes. It is understood that there is a high personal judgment about the
perception of water quality among farm households. Domestic use of irrigation water is highly
correlated with individual’s perception of the quality of their irrigation water and lack of
alternative sources.
24.17
9.93
15.56
50.33
Irrigation water Protected well/spring
Unprotected well/spring Carried from home
98
Figure 5-4: Major problems related to irrigation water sources
Source: Author’s computation using own survey data.
As shown in Table 5-6, the incidence of diseases by drinking water sources on irrigation field, the
highest proportion of incidence of diarrhea, other water-related illness and overall morbidity is
reported among households using irrigation water sources for drinking on the field. On the other
hand, households carrying water from home had better health status in the preceding two
months before the survey in terms of diarrhea incidence and overall morbidity status. This might
indicate that households who carry water from home have better health knowledge than others.
Moreover, households using protected wells had also better health status compared with
households using unprotected wells and/irrigation water except for diarrhea incidence.
Table 5-6: Incidence of diseases by drinking water source on irrigation site
Sources Diarrhea Water-related
diseases Morbidity Obs.
Irrigation water 14.12 17.06 43.24 340
Protected well 10.71 11.43 33.57 140
Unprotected well 10.24 13.78 36.61 254
Carried from home 7.35 10.82 28.85 721
Source: Author’s estimates using own survey data.
Using irrigation water for domestic purpose is not uncommon in many rural areas of Ethiopia
where improved drinking water is limited—particularly on-site access to clean water supply is
lacking. Figure 5-5 shows the number of households who use irrigation water for various types
of household purposes. Only 22 percent of households do not use irrigation water other than for
production purposes while more than 22 percent of households reported that they get their
145134
2919
0
25
50
75
100
125
150
Scarce water No problem Poor waterquality
Infection ormosquitoes
Nu
mb
er o
f h
ou
seh
old
s
99
drinking water from irrigation water sources. Most respondents indicated that irrigation water is
used mainly for livestock watering, gardening and washing/bathing purposes.
Figure 5-5: Irrigation water use for domestic purposes
Source: Author’s computation using own survey data.
As discussed earlier, most farmers have access to more than one water sources for irrigation and
they also use various irrigation methods which constrained us to further investigate if any
systematic pattern exists between households’ health status and irrigation water
sources/irrigation methods.
5.5.1.2 Estimation results: Impact of irrigation on water quantity and diarrhea
In this section, we try to examine the impacts of irrigated agriculture on health outcomes,
household income, demand for health inputs and household water quantity/quality using the full
sample. The impacts of irrigation on household water quantity, measured by l/p/c/d is presented
in the Appendix A (column 1—3, Table A2). The results show that irrigated households consume
relatively more water than non-irrigator households after controlling distance to drinking water
source, container size and other demographic variables (column 3, Table A2). The result further
suggests that being an irrigator household does not automatically translate into high per capita
water consumption. The influence is mediated via other channels (column 2, Table A2). In
addition, the marginal health gains associated with the increase in household water consumption
(lower than half a liter per day per person) would be imperceptible.
The health impacts of irrigation are analyzed both at the individual and household level. Table
5-7 presents the health effects of irrigation on non-preschool household members in the last two
months before the survey. Columns 1—4 present individual level estimated regression while the
178
139 136
75 69 67
0
50
100
150
200
Nu
mb
er o
f h
ou
seh
old
s
100
last two columns report the household level analysis. In column 5, the outcome variable is 1 if
there is any reported diarrhea cases in the household in the last two months preceding the survey
and 0 otherwise. In column 6, we counted the number of diarrhea cases per households.
As the result shows irrigation does not appear significantly impact diarrhea prevalence in the first
two columns of Table 5-7. The effect of irrigation on diarrhea, however, becomes large and
statistically significant in the subsequent regression results. This dramatic change happens when
we control the model for the proportion of irrigated land. This variable probably predicts the
extent to which the households are engaged in irrigated agriculture. On the other hand, the
dummy variable for irrigation could not capture the ‘pure’ influence of being irrigator household
on diarrhea incidence. The estimated coefficient of the irrigation dummy variable is
contaminated with kebele fixed effects, that is, the estimated coefficient partially captures kebele
fixed effects as irrigator and non-irrigator households are mutually exclusive at the kebele level.
Generally, the association between irrigation and diarrhea incidence appears robust in both
model specifications when we control for the proportion of irrigated land. For instance, a person
living in irrigator household is four percentage points higher in the probability of contracting
diarrhea in the last two month prior to the survey than a person living in non-irrigator households
(columns 3—4, Table 5-7). Moreover, in column 6, the estimated incidence-rate ratio for
irrigation is 1.74, implying a 74 percent increase in the number of diarrhea prevalence (1.74
times) among irrigator households per time unit (last two months preceding the survey) for given
covariate values. The impact of water quality is also robust and associated with 6 percentage
points decrease in the likelihood of contracting diarrhea (column 4), or 44 percent reduction in
the number of diarrheal illness in the last two months preceding the survey (column 6).
Regarding other control variables, household pit latrine marginally increases the number of
diarrhea cases in both model specifications. As expected, household size is significantly
associated with poor health status, but the magnitude of the estimated coefficient is low in the
individual level analysis (column 4, Table 5-7). Age is also a significant determinant of individual
health outcomes, that is, on average older household members are healthier than younger ones.
Moreover, distance to the nearest health center is associated with household health status.
While household assets and the proportion of irrigated land have a protective effect against
diarrhea, livestock holding slightly increase the incidence of diarrhea incidence, particularly at
the household level. The proportion of irrigation land may partly reflect the income effects, that
is, household income is positively related to the proportion of irrigated lands.
101
Table 5-7: Health effects of irrigation—Diarrhea
(1) (2) (3) (4) (5) (6)
VARIABLES Model Model Model Model Model Model
Irrigating households (dummy) 0.014 0.009 0.039** 0.044*** 0.095* 1.744***
(0.011) (0.011) (0.017) (0.016) (0.056) (0.297)
Water quality (1= no E.coli) -0.063*** -0.057*** -0.057*** -0.051*** -0.176*** 0.557***
(0.011) (0.011) (0.011) (0.011) (0.046) (0.087)
Water collection time (1=30min/less) 0.018 0.017 0.011 0.014 0.075 1.149
(0.013) (0.013) (0.014) (0.014) (0.049) (0.188)
Pit latrine (dummy) 0.012 0.023* 0.031** 0.032** 0.115*** 1.457***
(0.016) (0.012) (0.013) (0.013) (0.041) (0.181)
Age in years: reference (5—6)
7—9 years -0.193*** -0.188*** -0.190*** -0.188***
(0.043) (0.041) (0.041) (0.041)
10—15 years -0.286*** -0.282*** -0.282*** -0.281***
(0.037) (0.036) (0.036) (0.035)
16 years/older -0.310*** -0.306*** -0.307*** -0.305***
(0.035) (0.033) (0.033) (0.033)
Gender (male=1) -0.007 -0.008 -0.008 -0.008
(0.012) (0.012) (0.012) (0.012)
Average household member age -0.023*** 0.905***
(0.006) (0.021)
Household size 0.010** 0.010*** 0.010** 0.011*** 0.078*** 1.360***
(0.004) (0.004) (0.004) (0.004) (0.014) (0.055)
Household density 0.008** 0.005 0.005 0.003 0.023 1.006
(0.004) (0.004) (0.004) (0.004) (0.016) (0.038)
Highest education completed -0.009*** -0.008*** -0.008*** -0.008*** -0.028*** 0.908***
(0.002) (0.002) (0.002) (0.002) (0.008) (0.022)
Distance to health center 0.005** 0.004** 0.004** 0.013** 1.042**
(0.002) (0.002) (0.002) (0.005) (0.021)
Livestock holding 0.006 0.007* 0.028* 1.118**
(0.004) (0.004) (0.016) (0.050)
Log of assets value -0.016* -0.016* -0.059* 0.826**
(0.008) (0.008) (0.035) (0.078)
Proportion of irrigated land -0.075*** -0.077*** -0.258*** 0.324***
(0.027) (0.027) (0.088) (0.093)
Constant 1.532
(1.299)
Observation 2149 2149 2149 2149 454 454
Model Chi2 228.84 280.39 244.64 336.54 124.88 255.92
Model p-value 0.000 0.000 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1; Regression coefficients in average marginal effects for column 1—5 and in IRR (Incidence-rate ratios) for column 6. Dependent variable: Diarrhea prevalence among household members older than 5.
102
5.5.1.3 Estimation results: Domestic use of irrigation water and household water quality, quantity,
and health outcomes
This section is the central objective of this chapter and presents the estimated results of the
impacts of irrigation water use for domestic purpose on storage water quality and household
members’ health status. As all irrigator households do not use irrigation water for domestic
consumption, we need to focus on the sub-sample households to examine the impacts of
irrigation water use for domestic purpose on water quality and quantity and health outcomes.
Again, the health impacts are analyzed both at the individual and household level.
The impact of irrigation water use for domestic purposes such as drinking and cooking on POU
water quality is presented in the Appendix A (columns 4—5, Table A3). The results do not provide
a strong evidence that suggests domestic use of irrigation water considerably deteriorates the
microbiological quality of household storage water. The results further indicate that the distance
between a household and irrigation water source is rather a stronger predictor of storage water
quality than actual usage of irrigation water for domestic purposes. That is, the farther the
irrigation water source from home, the better the storage water quality. This is more intuitive,
but the nonsignificant relationship between domestic use of irrigation water and storage water
quality is surprising. It might be the case that one-time water quality test result could not actually
reflect the average storage water quality for a given household for a long time; consequently, the
effects of domestic use of irrigation water on storage water quality might not be captured.
Moreover, domestic use of irrigation water can be seasonal and/or households can use it only
during the field work. In the latter case, the water quality-testing results of storage water could
be hardly attributed to irrigation water uses. Generally, this shows the complexity of how the
various pathways play out in terms of their influence on domestic water quality and quantity and
irrigation infrastructure under various conditions.
Table 5-8 presents the effects of the domestic use of irrigation water on diarrhea incidence. In
column 1, 3 and 5, we controlled for POU water quality whereas, in the other columns, we did
not control for POU water quality. We assumed that the health effects of the domestic use of
irrigation water, if any, is through household storage water quality. That is, if domestic use of
irrigation water had significant health effects, we would expect the estimated coefficients
associated with it to be statistically significant or have larger effects on health outcome measures
whenever we do not control for POU water quality. The result presented in Table 5-8 suggests a
lack of evidence that domestic use of irrigation water exacerbates diarrheal illness. Neither does
the magnitude of its effect on health outcomes change much when we control for POU water
quality across all estimated models. On the other hand, the impact of domestic use of irrigation
water on other water-related illness and overall morbidity has a larger and significant effect size
at the household level across all model specifications (Table A6 & Table A7). However, again,
the effect of POU water quality is limited to diarrheal illness only.
103
Table 5-8: Effects of domestic use of irrigation water on diarrhea
Dependent variable: Diarrhea prevalence (%) among household members older than 5
PROBIT POISSON
VARIABLES (1) (2) (3) (4) (5) (6)
Irrigation water use (dummy) 0.005 0.013 -0.019 0.002 0.991 1.072
(0.018) (0.020) (0.063) (0.069) (0.170) (0.196)
Water quality (1=no E.coli) -0.057*** -0.212*** 0.532***
(0.014) (0.052) (0.111)
Water collection time (dummy) 0.016 0.011 0.073 0.054 1.183 1.120
(0.019) (0.021) (0.067) (0.071) (0.255) (0.253)
Pit latrine (dummy) 0.028* 0.028* 0.106** 0.101* 1.350** 1.374**
(0.015) (0.015) (0.054) (0.053) (0.201) (0.201)
Age in years: reference (5—6)
7—9 years -0.128** -0.125**
(0.051) (0.051)
10—15 years -0.209*** -0.207***
(0.041) (0.041)
16 years or older -0.230*** -0.227***
(0.036) (0.036)
Gender (male=1) -0.014 -0.013
(0.015) (0.015)
Average HH member age 0.012** 0.012** -0.023*** -0.024*** 0.902*** 0.900***
(0.005) (0.006) (0.007) (0.007) (0.025) (0.025)
Household size -0.001 0.001 0.081*** 0.077*** 1.361*** 1.356***
(0.005) (0.005) (0.017) (0.019) (0.073) (0.074)
Household density -0.008*** -0.009*** 0.001 0.012 0.953 0.976
(0.003) (0.003) (0.020) (0.020) (0.046) (0.047)
Highest education comp. 0.005 0.005 -0.027*** -0.029*** 0.923*** 0.913***
(0.003) (0.003) (0.009) (0.009) (0.027) (0.028)
Distance to health center 0.008 0.014*** 0.015* 0.016* 1.046 1.047
(0.005) (0.005) (0.009) (0.009) (0.033) (0.033)
Livestock holding -0.014 -0.019* 0.022 0.046** 1.130** 1.188***
(0.010) (0.010) (0.019) (0.019) (0.063) (0.065)
Log of assets value -0.079*** -0.093*** -0.056 -0.075* 0.825* 0.796*
(0.029) (0.031) (0.041) (0.043) (0.091) (0.096)
Proportion of irrigated land -0.128** -0.125** -0.265*** -0.318*** 0.321*** 0.281***
(0.051) (0.051) (0.090) (0.098) (0.095) (0.087)
Constant 3.166 3.234
(3.125) (3.255)
Observations 1455 1455 302 302 302 302
Model Chi2 92.73 83.96 190.44 124.06
Model p-value 0.000 0.000 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1 Coefficients in columns 1—4 are in average marginal effects while in columns 5—6 are in IRR.
104
Our results do not suggest that domestic use of irrigation water (for drinking and/or cooking)
exacerbates diarrhea incidence among non-preschool household members in the study areas.
Moreover, the effects of domestic use of irrigation water on diarrheal diseases could be
significantly lower where access to improved drinking water supply is low. Irrigation technologies,
however, generally result in poor household health in terms of water-related diseases. A study
in Ghana found that a headache, blurred vision, and nausea or vomiting illness symptoms were
higher among irrigator workers (Clarke et al. 1997). They argued that this negative health effect
comes from increased pollution due to pesticide use. This indicates that the quality and quantity
of drinking water can easily become affected through irrigation agricultural practices due to
increased use of pesticides and agrochemicals and inappropriate water withdrawals methods
(Eriksson 2012; Horgby & Larson 2013). Moreover, irrigation might serve as a vector-breeding
habitat which affects people’s health (Ersado 2005; Ghebreyesus et al. 1999). Although we did
not estimate the health production function for malaria, Table 5-1 clearly indicates that malaria
incidence is significantly higher among irrigator households.
Moreover, in addition to more sick time, the negative health impacts associated with irrigation
may increase household medical expense for health care. Table A9 shows estimated medical
expenditure functions controlling for irrigation and other few exogenous variables. Households
with irrigation technology spend significantly more money on medical expenses than households
without irrigation. This might be because irrigator households are generally less healthy than
non-irrigator households, which in turn increases their medical expenditures. In contrast,
Asayehegn et al. (2012) argued that healthy households are more likely to adopt irrigation
technologies, implying that households with irrigation technologies are generally healthier than
households without irrigation. On the other hand, high medical expenditure might be associated
with high household income, which is also associated with irrigation practices. But, this might not
be always true. Burney et al. (2010) found that health-care expenditure between irrigator and
non-irrigator households are statistically nonsignificant.
Another important link between domestic use of irrigation water and health outcome is through
increased water quantity and household’s time allocation. Hutton and Haller (2004) showed that
the time-saving associated with water supply and sanitation generates the greatest economic
benefits. Our results confirm a strong positive association between time spent for domestic water
collection and access to irrigation, implying irrigator households spend significantly less time than
non-irrigator households controlling for distance to drinking water source and other household
characteristics (Table A4). This indicates that availability of irrigation water considerably reduces
the burden of water collection which is mostly borne by women and children.
Climate change and unpredictable weather poses a challenge and aggravates the vulnerability of
farmers to food insecurity. Considering the key role of agriculture in the country’s economy,
expanding SSI can bring many benefits to farmers through several pathways. Small-scale
105
irrigation technology raises income through increasing production and productivity and reducing
risks induced by climate variability. In the study areas, most households with access to irrigation
produce more than once per year by extending production in the dry season, which provides
additional household income. Asayehegn et al. (2012) reported that farmers in Ethiopia with
access to irrigation plant crop multiple times per year compared to rain-fed production.
Moreover, productivity (gross revenue per hectare) is higher with access to irrigation (FAO 2000).
Small-scale irrigation may be instrumental in reducing poverty as incidence and depth of poverty
were found to be lower among households with irrigation technologies (Gebregziabher 2008).
Where rainfall is scarce and erratic, access to sufficient quantity and reliable water resource is
critical to enhancing agricultural production. In keeping with the ministry of water resources
(MoWR), the country could tap only 5 percent of its total water resources for irrigated agriculture
and other uses due to lack of technology and finance. The potential for expanding irrigated
agriculture in Ethiopia is still large. Of the estimated 3.7 to 4.3 million hectares of potentially
irrigable land, less than 10 percent of this area is currently irrigated, and more than half of this
area is irrigated by traditional practices (MoARD 2009). Given that agricultural households
constitute the vast majority of rural households in the country, expansion of SSI farming is
becoming the priority of the government to improve small-scale agricultural production. In its
Growth and Transformation Plan (GTP), the government is keen to expand the use of
groundwater for irrigation by supporting smallholder farmers in the adoption and use of private
hand-dug wells and appropriate water-lifting technologies. Such interventions may generate
higher returns if multiple-use of water is incorporated into the design to improve WASH and
agricultural production where access to improved domestic water supply is inadequate.
Our result highlights that irrigation significantly increases women’s agricultural work load as
irrigation requires more labor than rain-fed agriculture. On the other hand, it also substantially
reduces women’s time for water collection (Table A4). Furthermore, it revealed that the benefits
of the domestic use of irrigation water use may be more significant in terms of reducing women’s
water collection time and increasing the availability of domestic water consumption than its
negative health impacts via poor POU water quality. Although our results are robust to various
model specifications, they do not mean ‘causal’ relationships between the outcome and our
variables of interest. The variable irrigation or domestic use of irrigation water is more likely to
be endogenous due to measurement errors and/or unobserved household characteristics. Our
results, however, show strong correlations that, in statistical parlance, are too robust to ignore.
The study may also be contaminated by the long recall and under-report bias for any health
symptoms, particularly for diarrhea. As diarrheal disease is not considered as serious health
issues, it may go unreported especially among adults. In addition, all household members were
not available during the interview, and in most cases, the household head and primary caretaker
were responsible for responding the questionnaires.
106
5.6 Conclusions and Recommendations
Small-scale irrigation interventions can impact health and nutrition outcomes in several
pathways, including increased food availability (in quality and quantity), increased agricultural
productivity and production and thus household income from market sales. With irrigation,
households become less dependent on rain and can grow and sell high-value crops throughout
the year. On the other hand, poorly designed irrigation system has the potential to increase the
burdens of diseases through an increased incidence of vector-borne diseases, such as malaria,
and it also affects the quality and quantity of domestic water supply. Rural households often use
irrigation water for domestic purpose where access to improved source is inadequate. Similarly,
although the primary objective of increasing access to improved rural water supply is for human
consumption, rural water supplies are often used for a range of additional services such as water
for livestock, home gardens, and SSI activities for the production of crops and vegetables. Such
non-exclusive water supply (through multiple water use), however, presents both threats and
opportunities. Availability of irrigation water substantially reduces women’s time for domestic
water collection; nevertheless, irrigation provides another pathway for domestic water
contamination (both chemical and microbiological), as irrigation largely requires the application
of agro-chemicals.
Our results show that irrigation and household health are significantly associated though there
might be a self-selection problem to attribute the poor health outcomes to irrigation. We could
not find any evidence, however, to support our initial hypothesis that irrigation can increase
diarrhea incidence through its impact on domestic water quality. The impacts of irrigation on
domestic water quality may not be substantial to increase the risk of poor water quality to cause
diarrhea. On the other hand, irrigation reduces women’s time for domestic water collection and
increases domestic water availability. Moreover, although domestic use of irrigation water
generally deteriorates the health status of households, we could not find a significant difference
in terms of diarrhea incidence between the groups.
This shows that the linkage between rural water supply, sanitation practices, and agriculture is
complex. In this complexity, interventions might work best when they draw on a range of
approaches and disciplines rather than working on a single intervention. The situation may be
improved if rural water supplies can be constructed to accommodate a range of services in
addition to the basic domestic water supply to link with income generating activities, such as
livestock rearing, homestead irrigation, where only one water source is available for all the needs
of a community. Similarly, interventions in SSI water management systems should acknowledge
the multiple uses of water where access to improved domestic water availability is limited. If
multiple-use of water systems is designed by considering the human health and environmental
risks, it would help to optimize water availability and quality for domestic and productive use of
this limited resource. Moreover, promoting improved WASH from ‘farm to fork’ to reduce agro-
107
chemical and microbiological contamination as irrigation water can be a source of contamination.
Ensuring rural households do not use irrigation runoff for drinking is vital to avoid water-related
diseases. This needs to build behavioral changes for sustained improvements in water and
sanitation services and key hygiene practices. Lastly, SSI intervention should incorporate
nutrition and health aspects through training programs and awareness campaigns. This might
help to maximize the potential gains from SSI and to minimize the risk associated with it in terms
of nutrition and health. Generally, if the multiple-use of irrigation water is recognized where
water availability is limited, and both WATSAN and agriculture are managed in an integrated and
indirect support for improved health and nutrition outcomes, multiple benefits, such as improved
health and nutrition outcomes, might be achieved and trade-offs reduced.
108
6. Determinants of Child Nutrition Outcomes
6.1 Introduction
In developing countries, a significant percentage of children are suffering from malnutrition and
micronutrient deficiency. It is estimated that malnutrition alone is responsible for more than one-
third of child deaths globally (Black et al. 2008; UNICEF 2013). Ethiopia is not an exception in this
regard, and the problem of child undernutrition and substandard child growth attainment in the
country is among the highest in the world. Nationally, 40 percent of under-five children are
stunted (low height-for-age), which often used as an indicator of chronic malnutrition), and 21
percent of children are severely stunted (CSA 2014). The proportion of child malnutrition is even
much higher in rural areas.
Figure 6-1: Percentage of undernourished children and average dietary supply adequacy trend over time
Source: Author’s illustration using data from FAOSTAT 2015.
Figure 6-1 shows the percentage of under-five children who are underweight over time and
average dietary supply adequacy between 1992 and 2014. The figure indicates that the
percentage of under-five children who are underweight gradually reduced by less than a
percentage point per year between this period. Although child undernutrition has dramatically
declined over the last decade; the prevalence rate is however still unacceptably high.
10%
25%
40%
55%
70%
85%
100%
115%
1990 1994 1998 2002 2006 2010 2014
Average dietary supply adequacy
Prevalence of undernourishment - 3 years average
Percentage of children under 5 years of age who are underweight
109
The problem of chronic child malnutrition in Ethiopia has been there for many years, and the GoE
has recognized the gravity of the problem. To this end, several strategies have been devised by
the government to improve the overall nutritional and health status of the most vulnerable group
of the population, that is, infants and young children, and mothers. It launched various nutrition
and health programs to promote health education and environmental sanitation, provision of
micronutrients and treatment of extremely malnourished children (CSA & ICF International
2012). In 2008, the government launched the National Nutrition Programme (NNP) to address
the immediate and fundamental causes of malnutrition (UNICEF 2013). However, the
government should do much more if the prevalence of chronic malnutrition is to be substantially
reduced in the country.
The nutritional status of children is an important indicator of children’s health. It is the result of
complex interactions between feeding practice and the overall status of health and health-care
access and utilization (CSA & ICF International 2012). Several socioeconomic and cultural factors
influence child feeding practices and thereby the nutritional status of the child. The first two
years of life is understood to be important for children’s optimal growth and development.
Childhood illnesses, such as diarrhea, pneumonia, and micronutrient deficiencies, however,
impede on children’s path to optimal growth and development. The nutritional status of under-
five children can be evaluated using anthropometric data on height and weight. Based on this
evaluation children that are at increased risk of faltered growth can be identified.
Access to improved WATSAN is also important not only for child health and well-being but also
for child growth outcomes. A wide range of empirical evidence shows that lack of access to
improved WATSAN services is cause for poor child health (Bose 2009; Checkley et al. 2004; Esrey
1996; Günther & Fink 2010; Jalan & Ravallion 2003; Kumar & Vollmer 2013). The empirical
evidence base showing that lack of these services, especially sanitation, are causes of child
malnutrition is also growing and has received renewed attention in recent years.
Irrigated agriculture is also closely linked with nutrition and health and has long been considered
as a major driving force to fight acute malnutrition in developing countries (Kennedy & Bouis
1993). The link between irrigated agriculture and nutrition may seem intuitive and simple:
growing more and diversified crops provide households with more quality and quantity of foods
for a healthy and active lifestyle. A study of child nutrition in Kenya found evidence that there
are higher energy intakes and lower chronic malnutrition in children where there is access to
irrigation as compared to communities without access (Kirogo et al. 2007). The potential of
irrigation to improve nutrition and health outcomes, particularly in Africa, has been recently well-
documented (for a detailed review and discussion see, Domènech 2015; Domenech & Ringler
2013).
In line with the earlier discussion, the objective of this chapter is to assess the association
between drinking water quality and sanitation practices and child nutritional outcomes.
110
Identifying the major factors that influence child growth outcome where a significant portion of
the population is malnourished will help policy makers and practitioners to design the right
intervention to improve the public health at large.
6.2 WASH and Child Nutrition Outcomes
Unimproved drinking water and unsafe sanitation practices are common realities for a
substantial percentage of people in the developing countries. Empirical studies that examine the
relationship between WATSAN infrastructure and child nutritional outcomes are scarce and
existing knowledge is limited. Hebert (1985) shows that water quantity had a significant effect
on the growth and nutritional status of children above age three while water quality is more
important at younger ages. A study conducted in Southern Punjab, Pakistan provides supporting
evidence for this pattern. After controlling for potential confounding factors, the study finds a
strong correlation between quantity of water available in households and the prevalence of
diarrhea and stunting (van der Hoek et al. 2002).
Water supply and sanitation interventions can play a major role in fighting malnutrition, which is
an underlying cause of an estimated 2.6 million death of under-five children each year (Save the
Children 2012). Most of these children live in developing countries. More than 50 percent of the
chronic malnutrition among children in developing countries is due to repeated diarrhea and/or
intestinal infections as a result of unsafe drinking water, inadequate sanitation and insufficient
hygiene (Prüss-Üstün et al. 2008). Furthermore, episodes of repeated diarrhea due to inadequate
WASH and childhood undernutrition had significant negative long-term impacts on cognitive
development, and was linked with poorer school performance, lower adult economic
productivity, and increased mortality risk (Black et al. 2008; Humphrey 2009; Lorntz et al. 2006).
The causality between malnutrition and severe diarrheal diseases is still contentious in the
literature. Studies show that malnourished children are more susceptible to severe diarrhea
incidence and duration. On the other hand, repeated diarrhea can lead to malnutrition (Briend
1990; Checkley et al. 2008; Checkley et al. 2003; Dewey & Mayers 2011; Patwari 1999; Schorling
et al. 1990; Ulijaszek 1996). This suggests that diarrhea and malnutrition operate in a vicious
cycle. Reducing the prevalence of diarrhea and other infectious diseases has an impact on
nutritional status. That is interventions which improve WATSAN break the cycle and result in
better child nutrition directly and indirectly (Fenn et al. 2012; Guerrant et al. 2008; Lin et al. 2013;
Prüss-Üstün & Corvalán 2006; Spears 2013).
The WHO classified malnutrition into three broad categories based on the degree of severity and
the prevalence rate. Table 6-1 presents the WHO classification of malnutrition. Based on this
classification, the prevalence of child wasting is moderate although the prevalence of
underweight and stunting in Ethiopia can be termed as high and very high, respectively. Unless
111
the root causes of child malnutrition can be addressed timely, child malnutrition will remain a
great public health problem in the country.
Table 6-1: WHO classification of severity of malnutrition prevalence
Indicator Severity of malnutrition by prevalence ranges (%)
Low Medium High Very high
Stunting (low height-for-age) <20 20-29 30-39 >=40
Underweight (low weight-for-age) <10 10-19 20-29 >=30
Wasting (low weight-for-height) < 5 5-9 10-14 >=15
Source: WHO (n.d).
Trends in the nutritional status of children for the period 2000, 2005, 2011 and 2014 are shown
in Figure 6-2. It displays a declining trend in the proportion of children stunted, wasted and
underweight over the period. Stunting prevalence decreased by 31 percent (from 58% to 40%),
wasting reduced by 25 percent (from 12% to 9%), while underweight prevalence reduced by 39
percent (from 41% to 25%) over the last 15 years. Despite all the efforts made to address the
problem and the improvement achieved in the last two decades, the severity and prevalence of
chronic malnutrition is still quite high in the country.
Figure 6-2: Trends in nutritional status of under-five children, 2000-2014
Source: Ethiopian mini-demographic and health survey (CSA 2014).
58
12
41
51
12
33
44
10
29
40
9
25
0
10
20
30
40
50
60
Stunting Wasting Underweight
Un
der
un
utr
itio
n p
reva
len
ce (
%)
2000 2005 2011 2014
112
6.3 Data and Methods
For this analysis, we use the same primary household survey data described in chapter 2. In
addition to a wide range of demographic and health variables, the survey collected data on height
and weight of all under-five children to evaluate their nutritional status. Based on height and
weight measures, three indices of anthropometric indicators—weight-for-age, height-for-age,
and weight-for-height were calculated. A total of 565 under-five children were eligible to be
weighed and measured; however, height measurements for most children under six months old
were not obtained. As a result, data are presented for 547 of these children for weight-for-age
and 480 children for height-for-age, and weight-for-height measurement indices. The focus of
this analyses, therefore, is exclusively for children under-five years of old.
Anthropometric measurements are often used as indicators to measure under-five child
nutritional status. To form the standardized measures, weight and height/length measurements
converted to weight-for-age z-scores (WAZ), height-for-age z-scores (HAZ), and weight-for-height
z-scores (WHZ) for each child.34 Z-scores are generated using the new WHO child growth
standards using the perc Stata package (WHO Multicentre Growth Reference Study Group 2006).
The z-score is calculated as 𝑧𝑖 =𝑦𝑖−𝑦50
𝛿𝑦 where 𝑦𝑖 is the anthropometric outcome, 𝑦50 is the
median of the WHO reference population, and 𝛿𝑦 is the standard deviation of the WHO reference
population. Thez-scores indicate the number of standard deviations a child is below or above
what should be normal for his/her age in comparison to a reference child who was fed and raised
in an environment that favored optimal growth.
Measurement of nutritional outcomes
In our empirical analysis, we focus on two anthropometric measures: underweight and stunting
indicators of nutritional outcomes—which are often considered as a good indicator for child
nutrition outcomes. Child nutritional status is measured by weight-for-age and height-for-age z-
score. Underweight and stunting also measured as indicator variables: underweight (measured
by a dummy variable equal to 1 if the child’s WAZ score is less than -2, otherwise 0), and stunting
(measured by a dummy variable equal to 1 if the child’s HAZ score is less than -2, otherwise 0).
These indicator variables are measured as standard deviations from the median of the reference
population to describe whether a child was underweight or stunted. A weight-for-age z-score of
minus two standard deviations (-2 SD) from the reference population median is considered
underweight. A low weight-for-age (underweight) measurement can be used as an objective
34 The weight-for-height anthropometric indicator also used as to identify the proportion of population who are
overweight and obese. Children more than two standard deviations (+2 SD) above the median weight-for-height are considered overweight or obese. A recent report shows that a small proportion (about 2%) of children below age five years are classified as overweight or obese (+2 SD) in Ethiopia (CSA & ICF International 2012).
113
indicator for both form of malnutrition, that is, either chronic or acute malnutrition (Gray et al.
2006). On the other hand, a child with a low height-for-age measurement below the minus two
SD is regarded as stunted compared to the median of the reference population. The
anthropometric indicators of the nutritional status of children were calculated using new growth
standards published by the WHO in 2006 based on a new reference population collected in the
WHO Multicenter Growth Reference Study (2006). The new child growth standards show how
children grow under favorable environments all over the world, regardless of any social and
economic influences.
Control variables
As we are interested in examining the correlation between storage drinking water quality and
sanitation practice on child nutritional status, the key independent variables in this analysis are
drinking water quality and availability of a pit latrine, child stool disposal behavior, and irrigation
agriculture. In addition to child-specific and maternal characteristics, we also have controlled for
a range of household and village-level variables potentially correlated with both child
malnutrition and household’s WATSAN services. The variable delivery is a dummy variable for
whether a mother delivered her baby with the help of a health professional such as a doctor,
nurse or health extension worker. The variable antenatal care (ANC) indicates the number of
visits from a health professional during pregnancy period of her child. Dietary diversity score
measures the mean number of food groups (six food items) consumed in the last 24 hours before
the survey. Table 6-2 presents the descriptive statistics of the main variables for this analyses.
All the remaining variables are as defined in the preceding chapters.
Table 6-2: Descriptive statistics
Variables Obs. Mean SD
Prevalence of underweight 547 0.26 0.44
Prevalence of stunting 480 0.41 0.49
Prevalence of wasted 480 0.08 0.27
Weight-for-age Z score 547 -1.32 0.99
Height-for-age Z score 480 -1.70 1.20
Weight-for-height Z score 480 -0.62 0.98
Dietary diversity score (6 food items) 505 0.34 0.15
Delivery with health professionals 565 0.16 0.37
Number of antenatal care visits (ANC) 565 1.93 1.50
Source: Author’s estimates using own survey data.
6.4 Results and Discussion
Figure 6-3 shows the relationship between different anthropometric indicators. The graphs
indicate that height-for-age and weight-for-height are not correlated, whereas there tends to be
a positive correlation between height-for-age and weight-for-age. Besides, the distributions of
114
the z-scores for the three anthropometric indicators and the overall prevalence of
malnourishment are presented in Figure 6-4. The distribution of z-scores presented in Figure 6-4
provides strong visual evidence of malnourishment in all anthropometric indicators. It is easy to
see from the figures that the problems of low weight-for-age and height-for-age z-scores are
highly visible.
Figure 6-3: Correlation between different anthropometric indicators
Source: Author’s computation using own survey data.
-60
5
wa
z
-6 0 5haz
-60
5
wh
z
-6 0 5waz
-60
5
wh
z
-6 0 5haz
115
Figure 6-4: Distribution of z-scores for under-five children
Source: Author’s computation using own survey data.
Table 6-3 shows the percentage of under-five children in our sample classified as malnourished
based on the three anthropometric indices of nutritional status: height-for-age, weight-for-
height, and weight-for-age. The table shows that 29 percent of under-five children are
underweight (have low weight-for-age), 41 percent of children are stunted (low height-for-age),
and 8 percent of children are wasted (low weight and height). The results generally indicate that
child underweight and wasting is high while stunting is considered as very high in the study areas.
These results are also pretty much comparable to the rural national average prevalence rate,
which is estimated at 30.4 (underweight), 46.2 (stunted), and 10.2 (wasted) (CSA & ICF
International 2012, p. 159).
Table 6-3: The prevalence of malnutrition based on the WHO 2006 reference group
Variables Total Male Female
Obs. % Obs. % Obs. %
Underweight 547 26.32 252 29.36 295 23.73
Stunted 480 40.63 223 43.05 257 38.52
Wasted 480 8.13 223 8.97 257 7.39
Source: Author’s estimates using own survey data.
0.1
.2.3
.4
Density
-4 -2 0 2 4Z-score
WAZ
0.1
.2.3
.4
Density
-5 0 5Z-score
HAZ
0.1
.2.3
.4.5
Density
-4 -2 0 2 4Z-score
WHZ
Distribution of z-scores
116
Table 6-4 presents the prevalence of child malnutrition by household demographic
characteristics. The result shows that children older than 11 months are more likely to be
underweight and stunted. Also, there was a significant difference in child malnutrition by sex of
the child; boys were more likely to be underweight, stunting and wasted than girls. The
proportion of undernourished children varies by housing characteristics too. There is also a
marked difference in the prevalence of child malnutrition by districts. Children from Fogera
district are more likely to be undernourished than children from Mecha district. A potential
reason is that Fogera is often prone to flooding during the rainy season and open defecation is
more widespread, which can easily affect the water quality.
Table 6-4: Malnutrition prevalence in under-five children by demographic characteristics
Variables Obs. Underweight (%) Obs. Stunting (%) Wasting (%)
Age of child (months)
6 - 11 55 23.64 47 19..15 8.51
12 - 23 100 31.00 98 34.69 16.33
24 - 35 130 26.15 123 49.59 4.88
36 - 47 122 30.33 120 42.50 5.83
48 - 59 92 29.35 91 43.96 6.59
Sex of child
Male 295 29.37 223 43.05 8.97
Female 252 23.73 257 38.52 7.39
Housing roofing material
Corrugated iron sheet 499 26.05 435 41.15 8.05
Thatch 48 29.17 45 35.56 8.89
Districts
Fogera 347 28.24 307 41.69 10.10
Mecha 200 23.00 173 38.73 4.62
Source: Author’s estimates using own survey data.
The bivariate correlation between child nutritional status and household WATSAN characteristic
are also briefly discussed below. Table 6-5 presents the prevalence of malnourishment in under-
five children by WASH characteristics. The variation in child malnutrition based on improved or
unimproved water sources and types of drinking water sources is not systematic. However, the
prevalence of child malnutrition variation by household storage water quality is substantial.
Children from households whose storage water is contaminated with E.coli are more likely to
have inferior nutrition outcomes than children of households whose storage water is not
contaminated. A surprising result which comes out from Table 6-5 is that children of irrigating
households are more likely to be undernourished than children of non-irrigating households
across all anthropometric indicators. This finding is counter to what we hypothesized in the
117
earlier discussions. The common understanding is that irrigated agriculture is likely to improve
household food security and nutrition. Irrigation helps farmers to produce more and grow more
micronutrient foods. However, the environmental impact of irrigation infrastructures that
influence child health status may not be captured at the household level. On the other hand,
poor child nutritional outcome among irrigated households could be partly explained by low child
care due to high labor demand for irrigated agriculture. It might be also the case that the increase
in household income due to irrigated farming might not be spent on consumption that improves
child nutrition outcomes.
Table 6-5: Malnutrition prevalence in under-five children by WASH characteristics
Variables Obs. Underweight (%) Obs. Stunting (%) Wasting (%)
Water source based on WHO
Improved 267 27.34 235 37.45 8.94
Unimproved 280 25.36 245 43.67 7.35
Primary drinking water source
Private-protected dug well 30 33.33 26 46.15 7.69
Shared-protected well/spring 237 26.58 209 36.45 9.09
Unprotected well/spring 218 27.06 191 41.88 7.85
Surface water 62 19.35 54 50.00 5.56
Storage water quality
Contaminated 321 28.97 289 41.52 10.03
Uncontaminated 226 22.57 191 39.27 5.24
Handwashing with soap
Yes 152 29.61 128 39.06 9.38
No 395 25.06 352 41.19 7.67
Latrine
Yes 222 25.68 188 37.23 6.91
No 325 26.77 292 42.81 8.90
Irrigation farming
Yes 368 30.42 323 45.20 9.29
No 179 17.88 157 31.21 5.73
Source: Author’s estimates using own survey data.
The variation of child malnutrition can be influenced by household income level. Improved
household income could be associated with improved diet and better access to health services,
which in turn can develop children growth outcomes. As shown in Table 6-6, the correlation of
household income with child nutritional status is not much clear. We can see that children from
the top 40 percent of the income distribution are relatively better off, however.
118
Table 6-6: Malnutrition prevalence in under-five children by income distribution
Nutrition indicator
Income quintile Total
Poorest Poorer Average Richer Richest
Underweight (%) 30.84 25.23 26.36 25.93 23.42 26.33
Stunting (%) 41.57 44.33 42.11 35.71 39.60 40.63
Wasted (%) 8.99 11.34 11.34 7.14 4.95 8.13
Source: Author’s estimates using own survey data.
6.4.1 Multivariate regression results
Having investigated the bivariate correlation between storage water quality and sanitation
practices, and child nutritional status, this section employs multivariate regressions to control for
various confounding factors such as child and maternal-specific, and household and community-
level characteristics. The regression results for two anthropometric measures of interest, WAZ
(underweight) and HAZ (stunting) are reported in Table 6-7 and Table 6-8, respectively. The OLS
results are presented in columns 1—3, while estimated results for the probit models are
presented under columns 4—6 in each table. Furthermore, the dependent variable for the OLS
regression is the negative of the z-score such that a positive coefficient indicates a negative
correlation with weight/height. Robust standard errors adjusted for clustering at village level are
reported in each model. We start the estimation from the most parsimonious specification, i.e.
in the first column, the model is controlled only for the main variables of interest, and then we
showed progressively a more comprehensive specification. In this case, we can see how the
estimated coefficients remain stable to a full-blown model specification, which is our preferred
model. The explanatory power of the model is relatively modest for nutritional outcomes. Finally,
we discuss the results concurrently by each measure of anthropometric indicators next.
Results for weight-for-age z-score and underweight
The regression results presented in Table 6-7 indicate that storage water quality is positively
associated with weight-for-age z-scores, but this correlation is not statistically significant when
we controlled for health facility indicators (columns 3 and 6, Table 6-7). On the other hand,
though the sign is not expected, water collection time is robustly correlated with weight-for-age
z-score and underweight in all model specifications. We expected better child nutrition outcomes
when a household has access to a nearby drinking water source. Moreover, safe stool disposal is
statistically significant in the weight-for-age z-score but not in the preferred underweight model.
As expected, on average, it suggests that children living in villages with a low concentration of
child stools had improved weight-for-age z-score. In the parsimonious model, household’s pit
latrine is not strongly correlated with child nutrition outcomes, but in the subsequent model
specifications, pit latrine marginally associated with poor child nutritional status. This indicates
119
that pit latrine is associated with other control variables. For instance, this variable is moderately
correlated with distance to health facility—households are more likely to build latrines if they
live close to a health facility. It indicates that children living in households with a pit latrine are 8
percentage points increase in the likelihood of being underweight compared to children living
without a pit latrine (column 6, Table 6-7). This result may not be surprising as we have already
shown in the preceding chapter that pit latrines are not protective against diarrhea.
The other variables of interest, irrigation and livestock, are also significantly associated with child
nutrition outcomes, particularly in the underweight model (columns 4—6, Table 6-7).
Unpredictably, children from irrigating households are 13 percentage points more likely to be
underweight than children from non-irrigating households (column 6, Table 6-7). This can be
partly explained by lower child care due to high labor demand for irrigated agriculture (Komatsu
et al. 2015; Miller & Urdinola 2010). Normally, irrigated agriculture is more labor-intensive than
rain-fed agriculture (Namara et al. 2011). Moreover, irrigator households often produce twice a
year and demand more family labor than conventional farming households. On the other hand,
livestock ownership significantly improves child weight-for-age z-scores. Livestock improves child
nutritional outcomes by supplying richer animal proteins, such as cow milk, which may be the
only important source of animal protein and micronutrients for children in rural areas. Livestock,
however, may also be associated with poor sanitary environment.
Regarding other control variables, child age is negatively correlated with weight-for-age z-scores
(young children are less likely to be underweight than older children). Male children are slightly
more likely to be underweight than female children. The weight of a boy child is likely to be 0.14
z-scores lower than that of a girl (column 3, Table 6-7). Mother’s age, household size and level
of education appear not to have a significant correlation with child weight-for-age. As expected,
dietary diversity score and household assets are positively correlated with child nutritional status
in both model specifications.
In terms of health control variables, we find a strong positive association between weight-for-
age z-score and the number of antenatal care (ANC) visits. The effect of ANC appears large and
robust, and this could be that this variable may capture health knowledge of the primary
caretaker and household wealth. Also, children delivered by the help of a health professional are
15 percentage points lower in the likelihood of being underweight than children delivered
without the help of any health professionals. We could also not be surprised about the non-
significant influence of the variable distance to the health center, as this variable could not
capture the effects of quality of health services. Moreover, its effect may have been already
partially captured by the other health indicator variables. Because distance to health facility can
also influence the variable ANC and delivery.
120
Table 6-7: Multivariate regression model for weight-for-age/underweight
(1) (2) (3) (4) (5) (6)
VARIABLES OLS OLS OLS PROBIT PROBIT PROBIT
Water quality (1= no E.coli) -0.199* -0.183* -0.081 -0.104** -0.102** -0.069
(0.100) (0.097) (0.100) (0.043) (0.041) (0.042)
Water collection time (1=30min/less) 0.222* 0.190* 0.220** 0.153*** 0.155*** 0.142**
(0.115) (0.112) (0.106) (0.058) (0.056) (0.056)
Safe stool disposal (village mean) -0.901*** -0.911*** -0.612*** -0.164* -0.159* 0.015
(0.193) (0.190) (0.189) (0.091) (0.089) (0.108)
Pit latrine (dummy) 0.083 0.122 0.162* 0.058 0.067* 0.080**
(0.095) (0.090) (0.087) (0.038) (0.038) (0.036)
Irrigating households (dummy) 0.211** 0.164* 0.182* 0.120*** 0.112*** 0.128***
(0.105) (0.098) (0.092) (0.041) (0.039) (0.040)
Livestock holding -0.073** -0.055 -0.008 -0.060*** -0.055*** -0.027**
(0.032) (0.035) (0.037) (0.012) (0.013) (0.013)
Child age in months 0.044*** 0.006 0.013*** -0.005
(0.009) (0.010) (0.004) (0.004)
Child age square (/100) -0.054*** -0.006 -0.018*** 0.005
(0.015) (0.016) (0.006) (0.006)
Child is male 0.154* 0.136* 0.055 0.067**
(0.079) (0.074) (0.039) (0.030)
Mother age -0.010 -0.005 -0.006 -0.007**
(0.008) (0.008) (0.004) (0.003)
Highest education completed -0.032** -0.022 -0.005 0.001
(0.015) (0.014) (0.008) (0.008)
Household size 0.030 0.038 0.015 0.019
(0.033) (0.032) (0.016) (0.016)
Dietary diversity score -2.263*** -1.178***
(0.293) (0.133)
Log of assets value -0.110** -0.052***
(0.049) (0.018)
Delivery with health professional -0.080 -0.147***
(0.110) (0.052)
Number of ANC visits -0.117*** -0.052***
(0.026) (0.010)
Distance to health center 0.001 0.005
(0.013) (0.006)
Constant 1.589*** 1.029*** 3.043***
(0.171) (0.255) (0.448)
Observations 547 547 494 547 547 494
R-squared 0.07 0.14 0.27 0.07 0.09 0.30
Model F-stat/Chi2 7.127 6.700 14.16 33.13 41.75 136.74
Model p-value 0.000 0.000 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1, and Probit in average marginal effect Note: the dependent variable for the OLS regression is negative of the z-score.
121
Results for height-for-age z-score and stunting
The regression results presented in Table 6-8 indicate that there is no statistically significant
association between storage water quality and sanitation variables and child height-for-age
outcomes in our data. Studies show that environmental factors, such as access to improved
WATSAN, matter for child height (Dangour et al. 2013; Fenn et al. 2012; Kabubo-Mariara et al.
2009; Spears 2013). For instance, Fenn et al. (2012) showed that WASH interventions reduce the
prevalence of stunting by 12 percent in Ethiopia. Similarly, Lin et al. (2013) indicated that children
living in households with improved sanitation environment and proper hygiene are taller for their
age compared to children living in contaminated environment. However, others do not find a
statistically significant correlation between child nutritional outcomes and WATSAN (Kabubo-
Mariara et al. 2009). This might be because these variables sometimes may not measure well the
environmental quality that children grow up. Height-for-age (stunting) is a cumulative indicator
of slow physical growth and reflects long-term malnutrition (Glewwe et al. 2002). It is worth
pointing that household storage water quality can change frequently and cannot capture/predict
the actual water quality being consumed over a long time. Again, distance to drinking water
sources and irrigated agriculture are negatively associated with child height-for-age while dietary
diversity score is positively correlated.
Regarding the other control variables, child age and gender are negatively correlated with height-
for-age z-scores. For instance, the height of male children, on average, is 0.25 z-scores lower than
female children; older children are more likely to be stunted than younger ones. This is consistent
with some studies that show that girls have relatively better nutritional status than boys (CSA &
ICF International 2012; Kabubo-Mariara et al. 2009). Furthermore, the health-related control
variables, again the results are robust, statistically significant and fairly large nutritional gains to
the number of ANC visits and delivery by a health professional. Although access to or utilization
of health facility is limited in many rural areas of the country, a study by Headey (2014) shows
similar results that four or more ANC is strongly associated with better height-for-age z-score.
Generally, our results suggest that there is some degree of correlation between WATSAN and
child nutritional status, particularly the impact of water collection time is robust to any model
specifications. As discussed above, the negative impact of water collection time is not actually
expected. We showed, in the previous chapters, that water collection time is negatively
correlated with household drinking water quality and child diarrhea incidence. In the same line
of arguments, children living with households having closer water source point should have
better nutritional status than children living with households having a faraway water source. This
could however be partly explained by poor quality of drinking waters, that is, households have
to travel long distance to get access to improved water sources.
The impact of irrigation agriculture on child nutrition outcomes is also counterintuitive. Although
irrigation is expected to improve household food availability and dietary diversity, which in turn
122
improves nutritional outcomes of the households (Olney et al. 2009; von Braun et al. 1989), child
nutritional outcomes may not be improved. The positive impacts of irrigation on household
nutritional status may not occur automatically. Irrigation helps farmers not only to grow and sell
high-value crops throughout the year and increases their income but also to improve the food
base available to them. We also show that irrigator households report higher income from
agriculture than non-irrigator households, but we do not find a significant difference in terms of
dietary diversity score between irrigator and non-irrigator preschool children. As most of the
common staple foods in many communities do not provide sufficient micronutrient, such as
vitamins, minerals, for younger children, emphasizing dietary diversity and nutritional content of
the available food base may be vital to children’s growth and mental development.
The association between nutritional outcomes and health indicator variables are also robust
across all model specifications and show fairly large nutritional gains than improved WATSAN
services. In this analysis, we are only able to explain a small portion of the variations in child
nutrition outcomes. It is worth pointing that child nutrition outcome is a result of complex
interaction of different variables, and WATSAN is one of the many ingredients for improved child
growth outcomes.
The economic burden of malnutrition is estimated to be huge and its health effect is last long.
Malnourishment can impair proper cognitive and physical development, increase susceptibility
to infectious diseases, which in turn affects health outcomes, school performance, and
productivity. Investing in programs that improve child nutrition outcomes would more likely to
produce large returns in the long term, particularly in areas where the prevalence of malnutrition
is high as the case in rural Ethiopia.
This analysis may be suffered from omitted variables such as mother’s nutritional status and
child’s birth weight and order. For instance, a child’s birth weight is an important indicator of the
child’s nutritional status and child’s vulnerability to infectious diseases and survival. Children who
are small at birth are more likely to be malnourished than children who are average or larger in
size at birth (CSA & ICF International 2012). Birth weights however were not known for many
babies as most of them delivered at home and not weighted at birth. On the other hand, the
mother’s nutritional status is also a good predictor of her child’s nutritional outcomes. Empirical
studies show that children of thin mothers are more likely to be stunted than the children of
overweight mothers. Improving maternal nutrition is, therefore, vital for improving children’s
health and nutrition status.
123
Table 6-8: Multivariate regression model for height-for-age/stunting
(1) (2) (3) (4) (5) (6)
VARIABLES OLS OLS OLS PROBIT PROBIT PROBIT
Water quality (1= no E.coli) -0.123 -0.124 -0.064 -0.037 -0.037 -0.007
(0.129) (0.128) (0.116) (0.054) (0.054) (0.047)
Water collection time (1=30min/less) 0.640*** 0.601*** 0.611*** 0.174*** 0.162*** 0.142***
(0.148) (0.146) (0.126) (0.062) (0.063) (0.051)
Safe stool disposal (village mean) -0.758** -0.673** -0.388 -0.193 -0.155 -0.009
(0.287) (0.259) (0.266) (0.131) (0.125) (0.121)
Pit latrine (dummy) 0.075 0.080 0.119 -0.010 -0.011 -0.001
(0.111) (0.112) (0.109) (0.049) (0.047) (0.042)
Irrigating households (dummy) 0.251* 0.206 0.205 0.107** 0.086* 0.094**
(0.145) (0.137) (0.124) (0.053) (0.052) (0.040)
Livestock holding -0.053 -0.038 0.029 -0.021 -0.018 0.020
(0.047) (0.050) (0.049) (0.018) (0.019) (0.017)
Child age in months 0.058*** 0.046*** 0.019*** 0.013**
(0.015) (0.015) (0.007) (0.006)
Child age square (/100) -0.079*** -0.069*** -0.023** -0.018**
(0.024) (0.023) (0.011) (0.009)
Child is male 0.231** 0.249** 0.043 0.057*
(0.114) (0.102) (0.046) (0.034)
Mother age -0.020* -0.012 -0.010** -0.006
(0.012) (0.011) (0.004) (0.004)
Highest education completed -0.039** -0.024 -0.012 -0.003
(0.017) (0.016) (0.009) (0.008)
Household size 0.083* 0.053 0.038** 0.024
(0.044) (0.041) (0.016) (0.015)
Dietary diversity score -2.361*** -1.461***
(0.361) (0.154)
Log of assets value -0.014 -0.037
(0.071) (0.024)
Delivery with health professional -0.230* -0.105**
(0.134) (0.053)
Number of ANC visits -0.155*** -0.060***
(0.035) (0.013)
Distance to health center 0.005 -0.001
(0.014) (0.006)
Constant 1.450*** 0.665* 1.686***
(0.252) (0.379) (0.624)
Observations 480 480 476 480 480 476
R-squared 0.07 0.12 0.26 0.03 0.06 0.29
Model F-stat/Chi2 6.569 6.132 16.18 15.61 29.40 177.45
Model p-value 0.000 0.000 0.000 0.000 0.003 0.016
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1, and Probit in average marginal effect. Note: the dependent variable for the OLS regression is negative of the z-score.
124
6.5 Conclusion and Implications
Ethiopia has made a remarkable progress in reducing child undernutrition; however, a substantial
portion of the population is still malnourished, and the country remains one of the most
undernourished populations in the world. The country also lags behind the rest of the world in
terms of access to improved drinking water and adequate sanitation facilities (WHO/UNICEF
2015). This deficiency can contribute to increased child morbidity and mortality, poor nutritional
outcomes, lower social and economic development. Using a primary household survey data and
anthropometric measurements, this chapter investigates the degree of association between
WATSAN variables and nutritional status of preschool children in rural areas of Ethiopia.
This analysis contributes to the limited literature on child malnutrition and WATSAN services. The
results reveal that preschool children in rural Ethiopia have a lag in height-for-age and weight-
for-age relative to the WHO reference populations and support previous evidence. Indeed, our
results explain a small portion of the variations in child nutrition outcomes which is possibly
influenced by a range of interconnected variables and often difficult to control for and
disentangle their effects using cross-sectional data. To improve nutritional outcomes in rural
Ethiopia, it needs a great effort to improve access to and utilization of health service and
micronutrients in children’s diets on top of improving the WATSAN ladders. As the results
suggest, the quality and quantity of complementary feeding practices for children should be
promoted. Providing diverse food groups that supply all the vital micronutrients helps meeting
the requirements for optimal child growth and healthy lives. Furthermore, it is shown that
parental education is an important determinant of nutrition outcomes and can produce
reasonably large effects on child growth outcomes (Behrman & Wolfe 1987; Thomas et al. 1991);
nonetheless, as discussed earlier, both maternal and paternal levels of education are extremely
low in the study areas. In the long-term, maternal educational improvement could be an
important channel to improve child malnutrition.
125
7. General Conclusions and Policy Recommendations
In the developing world, lack of access to improved WATSAN services are major causes of
morbidity and mortality for large portions of the population. Expanding the provision of access
to improved WATSAN services is, therefore, a priority among government and development
partners in many developing countries. Increasing investments in WATSAN infrastructures and
promoting key hygiene practices are generally expected to improve the health, social and
economic benefits of the population—especially by reducing the risk of many communicable
diseases among young children.
Inadequate access to improved WATSAN services remains a major cause of health problems in
Ethiopia—particularly in the rural areas. Lack of clean drinking water and unsafe sanitation
practices are the main causes of diarrheal diseases among under-five children (CSA & ORC Macro
2006; Usman et al. 2016). The negative health impact of contaminated water is worsening
because most rural households obtain their drinking water from unprotected sources, and they
often consume the water without any prior treatment. Moreover, many rural households have
poor understanding of the importance of proper hygiene practices, which further increases the
risk of contracting infectious diseases. Household water sample tests also indicate that poor POU
water quality is a significant problem in the study areas. Besides the evidence that most of the
so-called ‘improved’ water sources in rural areas do not provide safe water for consumption, the
problem of lack of improved drinking water is compounded by POU water contamination due to
unsafe water storage and handling practices. Although access to clean water and pit latrines has
increased significantly during the last decade, many of the surveyed households did not regard
the progress as satisfactory in terms of access to clean water supply. Rural households
complained about a lack of access to improved water sources, poor water quality, and having to
travel long distances to access drinking water.
The GoE defines 15 liters of water per capita per day within 1.5 km of their home as a primary
indicator for access to improved rural water supply. However, this does not include water safety
at POU and water continuity throughout the year. As discussed earlier, most rural water supply
often contains water unsafe for drinking; in some instance, water sources do not distribute water
continuously throughout the year due to limited capacity and maintenance and repair problems
(MoW 2007; UNDP 2006; Beyene 2012). This implies that households may be forced to find out
alternative water sources which are of poor quality or less convenient and more distant sources.
As this stand-alone indicator conceals various inequalities, a broader approach should be
developed that take into account the basic pillars of access to improved water supply in rural
areas. While achievement on improved water supply coverage over the last decade has been
applauded, investments should be made to make sure that existing services are sustainable while
improving its quality.
126
According to the WHO/UNICEF (2015) JMP progress report on WATSAN over the last two decades
clearly indicated that sanitation coverage is lagging far behind water supply coverage. To meet
the SDG-6 which aims at ensuring access to WATSAN for everyone there should be more
concerted and coordinated actions. Unless efforts to increase access to improved WATSAN
services are intensified and implemented in conjunction with the promotion of proper hygiene
practices, communicable diseases will continue to remain a major cause of morbidity and
mortality in the country.
The study relied on a primary household survey data from rural areas in Ethiopia where water-
related diseases are quite prevalent health problems due to lack of access to clean water and
basic sanitation facilities. The household survey was conducted over a five month period
between February and June 2014. The survey collected a range of information on demographic
and socioeconomic characteristics including water quality sample testing at household’s water
storage and sources to determine the level of E.coli bacteria (CFU/100ml water). In addition,
anthropometric measurements, such as height and weight, were also collected for all under-five
children.
7.1 Potential Intervention Areas
The study suggests that the AG-WATSAN nexus requires a mix of instruments to address the
problem of health and nutrition outcomes in the study areas. That is, there is no one silver bullet
for the problem. It needs a more integrated cross-sectoral approach from various sectors such as
agriculture, WASH, health, nutrition, and education from different actors such central and local
government, and NGOs. Creating an enabling environment to facilitate the multi-sectoral
approach (i.e. working together by avoiding sectoral thinking about the problem) may bring long-
term and sustainable solutions for the problem. Based on our empirical findings, we highlight the
key policy implications coming from this work. We discussed four main potential intervention
areas, which can be targeted, to improve the existing poor WASH services and health and
nutritional status in the study areas, and we discussed them below one after the other.
7.1.1 Rural water supply
About 74 percent and 58 percent of the water sample from sources and household storage were
positive for E.coli bacteria, indicating that most rural population is at high risk of waterborne
diseases. Access to an improved drinking water source is low in the study areas—only 50 percent
of the surveyed households have access to protected or ‘improved’ water sources; however,
more than 43 percent of these households’ POU water was contaminated with fecal matters. The
findings indicate the rampant drinking water problems both at the POS and POU. The following
recommendations are made to bring immediate health benefits by addressing the problem of
poor POS and POU water quality.
127
Efforts should be increased to further expand the provision of improved drinking water supply to
increase access for the unserved rural population. Improving access to clean water supply not
only increases the quantity of clean water available for households but also allows households to
save time by reducing the distance between each household and the nearest water access point.
The available time from water collection may be used to look after the children and to engage in
other productive activities to generate income. The study also shows that shortening the distance
to water source points improves not only POU water quality but also health outcomes. However,
such infrastructure investments require many years of financing which might be only feasible in
the long-term for a country like Ethiopia. The following interventions can improve the overall
situation of poor water quality both at the POS and POU in the short-term. Such simple
interventions substantially reduce avoidable morbidity and mortality caused by lack of access to
improved WASH services.
Available water source points should be properly protected because protection of water source
points is a first-step to lessen the risk of contamination. Private-well water sources should not be
developed close to or downslope from household’s latrine to prevent potential seepage. Many
of the private-wells in the study areas are bucket-wells and they are often shallow and
inadequately protected and thus they can be contaminated easily by microorganisms and other
external pollutants such as latrine, animal droppings, dirty ropes and buckets and households
waste. Our results show a significant positive correlation between the presence of a pit latrine
and the POU water quality for households with own private-wells water source. Moreover, most
community water sources considered to be ‘improved’ showed the presence of E.coli which is
not in compliance with both the national and the WHO guideline standards.
Promoting household water treatment methods and products to address POU water quality
concerns. Given that most rural households obtain their drinking water from unprotected
sources, a simple household drinking water treatment practices substantially avoids the risk of
contaminated water. There is, however, a lack of awareness of the need to treat household
drinking water. POU water treatment on regular basis is almost negligible among the study
households because households’ perception regarding drinking water quality is that clean water
is ‘clean’ water. Promoting HWTS and the health risks of drinking contaminated water can,
therefore, bring significant progress in the provision of safe drinking water from source to mouth.
Increasing the provision of rural water supply alone may not be enough if households do not treat
their water or practice safe water storage and handling. The study suggests that uncontaminated
POU drinking water can reduce childhood diarrhea by 15 percentage points and up to 6
percentage points for adults. We, therefore, recommend all rural households to develop the
habits of HWTS to ensure safe water quality—particularly for drinking.
The study also indicates the importance of WUA in the provision of safe water supply in rural
areas. Households belong to a village with WUA have better storage water quality that who do
128
not. The provision of drinking water supply through community water scheme is the only existing
conventional approach to increasing access to clean water in many rural areas. Building the
capacity of WUA to enable them to repair and manage available water sources, and providing
training in water source protection and environmental sanitation is, therefore, critical in the
provision of sustainable water supply.
Establishing a water quality monitoring system which can monitor a set of common drinking
water quality indicators for rural water supply systems is essential to safeguarding public health.
Determining the public health risk associated with drinking water quality is useful; however, in
practice routine monitoring of pathogens is generally not carried out in many rural areas.
Community variations and household behavioral and sanitary factors are also key determinants
of household POU water quality. Unsafe sanitation practices, improper disposal or storage of
household wastes and poor environmental sanitation could be the primary causes of drinking
water contamination. Moreover, keeping domestic animals separately from household dwellings
and out of water source catchment areas can improve household water quality. Without proper
waste disposal and sanitation facilities, water source points are highly prone to gross
contamination from human and animal feces which are the primary sources of disease-causing
pathogens.
7.1.2 Sanitation infrastructure and practices
In terms of sanitation, we found that 42 percent the households have access to a simple pit latrine
while the rest of households defecate in the open. Such sanitation technology is considered
unimproved based on the commonly used WHO definition. In most cases, these latrines are not
fly-proof and do not have handwashing stations, squat-hole covers, and proper structures; and
become dysfunctional for many reasons, including the evidence that they are not connected to
any sewerage system and they can also easily collapse due to flooding.
Pit latrines should be therefore fly-proof equipped with a squat-hole cover and should be
hygienically maintained. The study reveals that availability of pit latrine increases the risk of
diarrhea incidence and the probability of a child being underweight. This indicates that
improperly build or poorly located and/or maintained latrines will not generate positive health.
Nevertheless, we are not encouraging open defecation per se, but rather arguing that existing pit
latrines are not good enough to achieve the desired health benefits of sanitation. Eliminating
human feces in the open through the construction of pit latrine alone is not enough to combat
diarrheal diseases if the oral-fecal transmission pathway is not completely blocked. Pit latrines
should be adopted if only adequate hygiene can be maintained, otherwise, they can pose a
serious health risk—especially if they are not fly-proof or insufficiently away from drinking water
sources.
129
7.1.3 Hygiene behavior awareness
Interventions targeting people’s behavior in WASH sector can be effective in the provision of safe
WATSAN services and in the fight against most infectious diseases. For instance, the provision of
improved water supply alone sometimes does not bring positive health gains primarily due to
household’s unsafe water storage and handling behavior which causes re-contamination at the
POU (Clasen 2015; Wright et al. 2004). Addressing the following points are instrumental in
improving individual and household level behaviors, which in turn improve health outcomes and
WASH environment.
Promoting handwashing with soap—a defensive mechanism for preventing germs from getting
into mouths and food, and avoiding diarrhea in young children. The results point towards the
importance of handwashing with soaps at critical times because unhygienic practices are key
pathways in the transmission of pathogens. As the study revealed, when primary caretakers
practice handwashing with soap at critical times, the risk of young children contracting diarrhea
was reduced by 6 percentage points. Proper hygiene practice is critical to realize the greatest
benefits of improved WATSAN services as well as to combat most communicable diseases. Yet,
again, inadequate water supply may cause households to limit their handwashing practices.
Promoting safe handling and disposal of child stools. The study also highlights that proper child
stool disposal behavior can reduce childhood diarrhea by 23 percentage points. However, the
behavior of safe child stool disposal is lacking in the study areas. Our findings also indicate that
household level variations alone do not appear to be appropriate to capture children’s potential
exposure to poor WATSAN services. Rather, the type of sanitation technologies and the practice
of child stools disposal at their neighbors are also important determinants of children’s health
outcomes. Direct contact with human feces can lead to diarrheal diseases. The safe disposal of
children’s feces is therefore tremendously important in preventing the spread of diseases.
However, most of the households perceived child feces to be less harmful than adult ones, and
child feces were therefore often left around or disposed close to a household’s living area where
children are often roaming and playing. This further highlights the lack of awareness among the
study households about the causes of diarrhea and the necessary remedial measures. Moreover,
many primary caretakers do not consider diarrhea to be a serious health problem as it is common
among young children. Education and public awareness campaigns could be an effective channel
to disseminate information that can improve child health associated with inadequate WASH
services. This is particularly important in many rural areas of the country, where mothers usually
have little education.
Promoting safe water vessels/containers. The study finds that 84 percent of the surveyed
households use jerrycan for water collection and storage; however, their storage water quality is
poorer than households using traditional clay vessel. Providing either safer storage containers or
promoting how to clean jerrycan properly would avoid substantial risk of water contamination.
130
Moreover, since adult women and girls are largely responsible for household water collection
and cleaning, targeting them in the provision of key hygiene education campaign will generate
substantial improvement in POU water quality.
On top of that, our findings indicate that WASH interventions are also required to make
households aware of the importance of safe WASH through educational campaigns so as to help
them change their long-held habits and hygiene behaviors. Primary caretakers often undermine
the critical role that good hygiene plays in improving overall health outcomes. Educating rural
communities on the potential sources of water contamination, proper water treatment methods,
safe disposal of child feces away from the domestic environment, and good hygiene practices
(for instance; handwashing with soap at critical times such as before preparing and eating foods,
after defecation and/or disposing of child stools) can result in significant health gains to the rural
population. Households should also be encouraged to have appropriate storage containers
where water continuity is not guaranteed or water is scarce but at the same time, containers
used for collecting/storing water need to be cleaned regularly using detergents to ensure safe
water quality.
7.1.4 Small-scale irrigation agriculture
Irrigation water is used for various domestic purposes in addition to its prime purpose of
agricultural uses. Domestic use of irrigation water includes drinking, cooking, washing and
bathing, livestock and aquaculture. This multiple-use of irrigation water is not planned by design
in most contexts, however. In rural areas where improved water supply is inadequate, drawing
drinking water directly from irrigation sources might serve to meet adequate water for domestic
needs. In such an environment, domestic and productive water uses are often not exclusive to
each other and the quality of the water is compromised as irrigation water can be a source of
drinking water contamination. An integrated effort on drinking water supply and irrigation water
management may bring an impact on the health and quality of life of those it benefits. Our result
indicates that although irrigation water use for domestic purposes does not deteriorate the
microbial quality of stored drinking water, it generally results in poor health outcomes.
Addressing the following concerns will be instrumental in maximizing the benefits of multiple-
use of irrigation water.
Constructing rural water supplies to accommodate a range of services to link with irrigated
agriculture in addition to the basic domestic water supply. Similarly, interventions in SSI should
acknowledge the multiple-use of water where improved domestic water availability is limited. If
multiple-use of water systems are designed by considering the health and environmental risks, it
would help to optimize the limited water resource availability and quality for domestic as well as
for productive uses. This further encourages a more integrated, cross-sectoral approach that
could deliver a greater and more sustainable impact of any intervention that targets health and
131
nutrition in rural areas. However, more research may be needed to generate the costs and
benefits of various intervention options and to identify the right technical and institutional
arrangements under which agriculture-WATSAN nexus could work to exploit the opportunities
opened up to change and improve human health and livelihood of rural households at large. As
a caveat, we should also be more careful that cross-sectoral approach might even worsen the
situation if the right institutional setup is not in place.
Promoting improved WASH services from ‘farm to fork’ to reduce agro-chemicals and
microbiological contamination as irrigation water can be a source of contamination. Most
households lack awareness about the detrimental health impacts of poor water quality, unsafe
sanitation, and inadequate hygiene practices. Promoting improved WASH in the field and
ensuring that households do not use irrigation runoff for drinking without appropriate POU water
treatment is, therefore, vital to avoid water-related diseases. Moreover, small-scale irrigation
promotion should incorporate nutrition and health aspects through training programs and
awareness campaigns. This helps to maximize the potential gains from SSI and to minimize the
risks associated with it in terms of nutrition and health.
Generally, agriculture, WASH, and nutrition are closely interlinked and complex. These complex
interactions are inextricably linked and in such a complex environment a single intervention
approach will not work. For instance, providing safe drinking water without safe stool disposal
and key behavioral changes will not work. Furthermore, improving nutrition without safe WASH
environment, then the latter will undo much of the good gained from improved diets. This
complexity, therefore, requires a mix of instruments needed to end hunger, to achieve food
security, to improve nutrition, to promote sustainable agriculture (SDG-2) as well as to ensure
access to water and sanitation for all (SDG-6) in the next decade. However, achieving the SDG-2
and SDG-6 targets in the next decade will require large economic resources, new-cost effective
and sustainable technologies and engagement of various actors at different levels.
Lastly, most SSI systems, such as shallow wells, tend to be developed and managed by farmers
using their own knowledge and skills. Providing technical and/or financial supports for such
households who desire to develop or improve their own groundwater supply could be an
alternative policy approach to link rural water supply and SSI, which in turn increases access to
improved rural water supply.
7.2 Limitations of the Study and Future Research Directions
One of the main limitations of the study is our reliance on the self-reported prevalence of
diarrhea and other illness symptoms, which are not clinically verified for our data. The study
would have given more insights if actual medical records were collected from the districts health
centers/posts. However, reliable morbidity data for such communicable diseases are difficult to
obtain from local health centers/posts due to lack of adequate health-care services in rural areas.
132
The study would have been also benefited if water sample quality tests were repeated for
certainty. Although one-time sampling information is useful, high level of E.coli may be a one-
time event occurrence, which does not allow us to capture seasonal impacts on groundwater
quality. Since the sampled households entirely rely on non-piped water supply sources, seasonal
changes could likely affect water quality in the household, which could also influence the level of
water quality measured. Conducting subsequent water sample testing over time can provide a
more representative water quality indicator. The study could have been also improved if water
samples had been taken from all possible community water sources and matching with
household water sample test results.
The other possible drawback of this study is that we did not look into the gender implications of
irrigation and the roles of women in irrigated agriculture. It is hypothesized that “improving
women’s access to and ownership of irrigation technologies and control over irrigated produce
would have a positive multiplying effect on nutrition and health outcomes.” Integrating gender
into future policy designs is essential to enhance the full potential impacts of irrigation
interventions on nutrition and health. An ongoing project in Ethiopia, Ghana, and Tanzania: “The
Feed the Future Innovation Lab for Small-Scale Irrigation and the Impact of Irrigation on
Agricultural Productivity, Nutrition, Health and Women’s Empowerment” has been working on
the gender dimension of impacts of irrigation on health and nutrition outcomes.35 We hope it
would shed some light on this regard.
In summary, we acknowledged the limitations of analyzing causal studies using cross-sectional
data; consequently, drawing such critical policy implications on the basis of it may be questioned.
However, in statistical parlance, the empirical results are too robust to ignore due to strong
correlations. Future studies may consider using a randomized control-trial (RCT) or panel dataset
to address this limitation. We also recommend future research to focus on assessing seasonal
changes in rural water supply and how it impacts storage water quality under multiple-use water
systems. Furthermore, it should focus on exploring on how individual level of behaviors related
to WASH affects both POS and POU water quality. Most importantly, future studies should look
at what kind of local institutional arrangements are more appropriate to facilitate an integrated
multi-sectoral approach to maximize the synergies and minimize the trade-offs in the nexus
among WASH-agriculture and nutrition.
35 The Feed the Future Innovation Lab for Small- Scale Irrigation, a project of the U.S. Government’s Feed the Future
Initiative, is a five- year project in Ethiopia, Ghana and Tanzania aimed at benefiting the region’s farmers by improving effective use of scarce water supplies through interventions in small-scale irrigation.
133
8. References
ALAM, N., WOJTYNIAK, B., HENRY, F. J., & RAHAMAN, M. M. (1989). Mothers' Personal and Domestic Hygiene and Diarrhoea Incidence in Young Children in Rural Bangladesh. International journal of epidemiology, 18(1), 242-247. doi: 10.1093/ije/18.1.242
Alderman, H., Chiappori, P.-A., Haddad, L., Hoddinott, J., & Kanbur, R. (1995). Unitary versus Collective Models of the Household: Is It Time to Shift the Burden of Proof? The World Bank Research Observer, 10(1), 1-19. doi: 10.1093/wbro/10.1.1
Amacher, G. S., Ersado, L., Grebner, D. L., & Hyde, W. F. (2004). Disease, Microdams and Natural Resources in Tigray, Ethiopia: Impacts on Productivity and Labour Supplies. The Journal of Development Studies, 40(6), 122-145. doi: 10.1080/0022038042000233830
Amenu, K., Spengler, M., André, M., & Valle Zárate, A. (2014). Microbial Quality of Water in Rural Households of Ethiopia: Implications for Milk Safety and Public Health. Journal of Health, Population, and Nutrition, 32(2), 190-197.
Arnold, B. F., & Colford, J. M. (2007). Treating water with chlorine at point-of-use to improve water quality and reduce child diarrhea in developing countries: a systematic review and meta-analysis. The American journal of tropical medicine and hygiene, 76(2), 354-364.
Asayehegn, K. (2012). Negative impact of small-scale irrigation schemes: A case study of Central Tigray Regional state, Ethiopia. Agricultural Research and Review, 1(3), 80-85.
Asayehegn, K., Yirga, C., & Rajan, S. (2012). Effect of Small-Scale Irrigation on the Income of Rural Farm Households: The Case of Laelay Maichew District, Central Tigray, Ethiopia. Journal of Agricultural Sciences, 7(1), 43-57. doi: http://dx.doi.org/10.4038/jas.v7i1.4066
Asenso-Okyere, K., Asante, F. A., Tarekegn, Jifar, & Andam, K. S. (2012). Addressing the Links Among Agriculture, Malaria, and Development in Africa In Shenggen Fan & R. Pandya-Lorch (Eds.), Reshaping Agriculture for Nutrition and Health (pp. 129). Washington, D.C.: International Food Policy Research Institute (IFPRI).
Atnafu, M. (2006). Assessment of Bacteriological Quality of Drinking Water Supply at the sources and Point of Use at Home in Worebabu District, South Wollo. M. Sc Thesis. Addis Ababa University, Addis Ababa. 34-57.
Bain, R., Cronk, R., Wright, J., Yang, H., Slaymaker, T., & Bartram, J. (2014). Fecal Contamination of Drinking-Water in Low- and Middle-Income Countries: A Systematic Review and Meta-Analysis. PLoS Med, 11(5), e1001644. doi: 10.1371/journal.pmed.1001644
Banks, D., Karnachuk, O. V., Parnachev, V. P., Holden, W., & Frengstad, B. (2002). Groundwater Contamination from Rural Pit Latrines: Examples from Siberia and Kosova. Water and Environment Journal, 16(2), 147-152. doi: 10.1111/j.1747-6593.2002.tb00386.x
Bartram, J., & Cairncross, S. (2010). Hygiene, Sanitation, and Water: Forgotten Foundations of Health. PLoS Med, 7(11), e1000367. doi: 10.1371/journal.pmed.1000367
Becker, G. S. (1965). A Theory of the Allocation of Time. The Economic Journal, 75(299), 493-517. doi: 10.2307/2228949
Begashaw, G. (2003). Community health, water supply and sanitation. Integrated water and land management research and capacity building priorities for Ethiopia, 98.
134
Behrman, J. R., & Deolalikar, A. B. (1988). Health and Nutrition. In C. Hollis & T. N. Srinivasan (Eds.), Handbook of Development Economics (Vol. 1, pp. 631-711): Elsevier.
Behrman, J. R., & Wolfe, B. L. (1987). How does mother's schooling affect family health, nutrition, medical care usage, and household sanitation? Journal of Econometrics, 36(1–2), 185-204. doi: http://dx.doi.org/10.1016/0304-4076(87)90049-2
Beyene, H. A. (2012). Factors Affecting the Sustainability of Rural Water Supply Systems: The Case of Mecha Woreda, Amhara Region, Ethiopia. Cornell University.
Bickenbach, J. E. (2008). Measuring health: The disability critique revisited. In 3rd annual international conference on ethical issues in the measurement of health and the global burden of disease at Harvard University, Cambridge, MA.
Billig, P., Bendahmane, D., & Smindale, A. (1999). Water and Sanitation Indicators Measurment Guide Academy for Educational Development, FANTA, Washington, DC.
Black, R. E., Allen, L. H., Bhutta, Z. A., Caulfield, L. E., de Onis, M., Ezzati, M., . . . Rivera, J. (2008). Maternal and child undernutrition: global and regional exposures and health consequences. The Lancet, 371(9608), 243-260. doi: http://dx.doi.org/10.1016/S0140-6736(07)61690-0
Blackden, C. M., & Wodon, Q. (2006). Gender, time use, and poverty in sub-Saharan Africa (Vol. 73): World Bank Publications.
Boelee, E., Laamrani, H., & van der Hoek, W. (2007). Multiple use of irrigation water for improved health in dry regions of Africa and South Asia. Irrigation and Drainage, 56(1), 43-51. doi: 10.1002/ird.287
Bosch, C., Hommann, K., Rubio, G., Sadoff, C., & Travers, L. (2002). Water and Sanitation. Chapter 23 In A sourcebook for Poverty Reduction Strategies (Vol. 2, pp. 371-404). Washington, D.C.: The World Bank.
Bose, R. (2009). The impact of Water Supply and Sanitation interventions on child health: evidence from DHS surveys: Mimeo, New Delhi: International Initiative for Impact Evaluation.
Bradley, D. (1977). Health Aspects of Water Supplies in Tropical Countries in Water, Waste and Health in Hot Climates. Richard Feacham, Michael Mc Garry and Duncan Mara (Eds.). London, J: Wiley and sons Ltd.
Braind, A., Nauges, C., Strand, J., & Travers, M. (2010). The impact of tap connection on water use: the case of household water consumption in Dakar, Senegal. Environment and Development Economics, 15(01), 107-126. doi: doi:10.1017/S1355770X09990076
Brick, T., Primrose, B., Chandrasekhar, R., Roy, S., Muliyil, J., & Kang, G. (2004). Water contamination in urban south India: household storage practices and their implications for water safety and enteric infections. International Journal of Hygiene and Environmental Health, 207(5), 473-480.
Briend, A. (1990). Is diarrhoea a major cause of malnutrition among the under-fives in developing countries? A review of available evidence. European journal of clinical nutrition, 44(9), 611-628.
Brocklehurst, C. (2011). Water, Sanitation and Hygiene: Foundations of Development The Global Water Crisis: Addressing an Urgent Security Issues [High-Level Expert Group Meeting]. Toronto, Canada: InterActionCouncil.
135
Bukenya, G. B., & Nwokolo, N. (1991). Compound Hygiene, Presence of Standpipe and the Risk of Childhood Diarrhoea in an Urban Settlement of Papua New Guinea. International journal of epidemiology, 20(2), 534-539. doi: 10.1093/ije/20.2.534
Burney, J., Naylor, R. L., & Postel, S. L. (2013). The case for distributed irrigation as a development priority in sub-Saharan Africa. Proceedings of the National Academy of Sciences, 110(31), 12513-12517.
Burney, J., Woltering, L., Burke, M., Naylor, R., & Pasternak, D. (2010). Solar-powered drip irrigation enhances food security in the Sudano–Sahel. Proceedings of the National Academy of Sciences, 107(5), 1848-1853.
Butterworth, J., Sutton, S., & Mekonta, L. (2013). Self-supply as a complementary water services delivery model in Ethiopia. Water Alternatives, 6(3), 405.
Cairncross, S. (1987). Water, Women and Children. In J. Pickford & B. Leedham (Eds.), The Benefit of Water Supply: Developing World Water II, London: Grosvener.
Cairncross, S. (1997). More water: better health. People and the Planet, 6(3), 10-11.
Cairncross, S., & Cuff, J. L. (1987). Water use and health in Mueda, Mozambique. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81(1), 51-54. doi: http://dx.doi.org/10.1016/0035-9203(87)90280-X
Cairncross, S., Hunt, C., Boisson, S., Bostoen, K., Curtis, V., Fung, I. C., & Schmidt, W.-P. (2010). Water, sanitation and hygiene for the prevention of diarrhoea. International journal of epidemiology, 39(suppl 1), i193-i205. doi: 10.1093/ije/dyq035
Cameron, J., Hunter, P., Jagals, P., Pond, K. (2011). Valuing water, valuing livelihoods: Guidance on social cost-benefit analysis of drinking-water interventions, with special reference to small community water supplies. World Health Organization.
Cameron, L. (2009). Does" improved" Sanitation Make Children Healthier?: Household Pit Latrines and Child Health in Rural Ethiopia (Vol. Working Paper No.42): University of Oxford. Department of international development. Young lives.
Cattaneo, M. D., Galiani, S., Gertler, P. J., Martinez, S., & Titiunik, R. (2009). Housing, Health, and Happiness. American Economic Journal: Economic Policy, 1(1), 75-105.
Checkley, W., Buckley, G., Gilman, R. H., Assis, A. M., Guerrant, R. L., Morris, S. S., . . . Network, I. (2008). Multi-country analysis of the effects of diarrhoea on childhood stunting. International journal of epidemiology, 37(4), 816-830. doi: 10.1093/ije/dyn099
Checkley, W., Epstein, L. D., Gilman, R. H., Cabrera, L., & Black, R. E. (2003). Effects of Acute Diarrhea on Linear Growth in Peruvian Children. American Journal of Epidemiology, 157(2), 166-175. doi: 10.1093/aje/kwf179
Checkley, W., Gilman, R. H., Black, R. E., Epstein, L. D., Cabrera, L., Sterling, C. R., & Moulton, L. H. (2004). Effect of water and sanitation on childhood health in a poor Peruvian peri-urban community. The Lancet, 363(9403), 112-118. doi: http://dx.doi.org/10.1016/S0140-6736(03)15261-0
Clarke, E., Levy, L., Spurgeon, A., & Calvert, I. (1997). The problems associated with pesticide use by irrigation workers in Ghana. Occupational Medicine, 47(5), 301-308.
Clasen, T. (2015). Household Water Treatment and Safe Storage to Prevent Diarrheal Disease in Developing Countries. Current Environmental Health Reports, 1-6. doi: 10.1007/s40572-014-0033-9
136
Clasen, T., & Bastable, A. (2003). Faecal contamination of drinking water during collection and household storage: the need to extend protection to the point of use. J Water Health, 1, 109-115.
Clasen, T., Cairncross, S., Haller, L., Bartram, J., & Walker, D. (2007). Cost-effectiveness of water quality interventions for preventing diarrhoeal disease in developing countries. Journal of water and health, 5(4), 599-608.
Crampton, J., & Aid, W. (2005). Maintaining clean water: contamination during water collection and storage in Addis Ababa.
CSA. (2003). Report on area and production of crops (Vol. Part II-B). Addis Ababa, Ethiopia.
CSA. (2013). Population projection of Ethiopia for all regions at wereda level from 2014 – 2017. Addis Ababa, Ethiopia.
CSA. (2014). Ethiopian mini demographic and health survey 2014. Addis Ababa: CSA.
CSA and ICF International. (2012). Ethiopia Demographic and Health Survey 2011. Addis Ababa, Ethiopia and Calverton, Maryland, USA: Central Statistical Agency (CSA) and ICF International.
CSA and ORC Macro. (2006). Ethiopia Demographic and Health Survey 2005. Addis Ababa, Ethiopia and Calverton, Maryland, USA.
Curtis, V., & Cairncross, S. (2003). Effect of washing hands with soap on diarrhoea risk in the community: a systematic review. The Lancet infectious diseases, 3(5), 275-281. doi: http://dx.doi.org/10.1016/S1473-3099(03)00606-6
Curtis, V., Cairncross, S., & Yonli, R. (2000). Review: Domestic hygiene and diarrhoea – pinpointing the problem. Tropical Medicine & International Health, 5(1), 22-32. doi: 10.1046/j.1365-3156.2000.00512.x
Curtis, V., Kanki, B., Mertens, T., Traore, E., Diallo, I., Tall, F., & Cousens, S. (1995). Potties, pits and pipes: Explaining hygiene behaviour in Burkina Faso. Social Science & Medicine, 41(3), 383-393. doi: http://dx.doi.org/10.1016/0277-9536(94)00341-P
Dangour, A. D., Watson, L., Cumming, O., Boisson, S., Che, Y., Velleman, Y., . . . Uauy, R. (2013). Interventions to improve water quality and supply, sanitation and hygiene practices, and their effects on the nutritional status of children. Cochrane Database Syst Rev, 8.
Dewey, K. G., & Mayers, D. R. (2011). Early Child Growth: How Do Nutrition and Infection Interact? Maternal & Child Nutrition, 7, 129-142. doi: 10.1111/j.1740-8709.2011.00357.x
Domènech, L. (2015). Is Reliable Water Access the Solution to Undernutrition? A Review of the Potential of Irrigation to Solve Nutrition and Gender Gaps in Africa South of the Sahara. IFPRI Discussion Paper 01428.
Domenech, L., & Ringler, C. (2013). The Impact of Irrigation on Nutrition, Health, and Gender: A Review Paper with Insights for Africa south of the Sahara. IFPRI Discussion Paper 01259. Available at SSRN: http://ssrn.com/abstract=2249812.
Dufour, A., Bartram, J., Bos, R., & Gannon, V. (2012). Animal Waste, Water Quality and Human Health. London: New York: International Water Association (IWA) and World Health Organization (WHO).
Edberg, S. C., Rice, E. W., Karlin, R. J., & Allen, M. J. (2000). Escherichia coli: the best biological drinking water indicator for public health protection. Journal of Applied Microbiology, 88(S1), 106S-116S. doi: 10.1111/j.1365-2672.2000.tb05338.x
137
Eriksson, S. (2012). Water quality in the Koga Irrigation Project, Ethiopia: A snapshot of general quality parameters.
Ersado, L. (2005). Small-scale irrigation dams, agricultural production, and health: theory and evidence from Ethiopia. World Bank Policy Research Working Paper(3494).
Esrey, S. A. (1996). Water, Waste, and Well-Being: A Multicountry Study. American Journal of Epidemiology, 143(6), 608-623.
Esrey, S. A., Potash, J. B., Roberts, L., & Shiff, C. (1990). Health benefits from improvements in water supply and sanitation: survey and analysis of the literature on selected diseases. WASH Technical Report. United States Agency for International Development(66).
Esrey, S. A., Potash, J. B., Roberts, L., & Shiff, C. (1991). Effects of improved water supply and sanitation on ascariasis, diarrhoea, dracunculiasis, hookworm infection, schistosomiasis, and trachoma. Bulletin of the World Health Organization, 69(5), 609.
FAO. (2000). Socio-economic impact of smallholder irrigation development in Zimbabwe: Case studies of ten irrigation schemes. FAO subregional office for East and Southren Africa. Harare, Zimbabwe.
Faurès, J.-M., Santini, G., & de Villemarceau, A. N. (2008). A livelihood approach to water interventions in rural areas and implications for Multiple Use Systems. In J. Butterworth, M. Keitjizer, I. Smout & F. Hagos (Eds.), Multiple-use Water Services. Proceedings of an international symposium held in Addis Ababa, Ethiopia. .
Fenn, B., Bulti, A. T., Nduna, T., Duffield, A., & Watson, F. (2012). An evaluation of an operations research project to reduce childhood stunting in a food-insecure area in Ethiopia. Public Health Nutrition, 15(09), 1746-1754. doi: doi:10.1017/S1368980012001115
Fewtrell, L., Kaufmann, R. B., Kay, D., Enanoria, W., Haller, L., & Colford, J. M. (2005). Water, Sanitation, and Hygiene Interventions to Reduce Diarrhoea in Less Developed Countries: A Systematic Review and Meta-Analysis. The Lancet infectious diseases, 5(1), 42-52.
Fewtrell, L., Kaufmann, R. B., Kay, D., Enanoria, W., Haller, L., & Colford Jr, J. M. (2005). Water, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis. The Lancet Infectious Diseases, 5(1), 42-52. doi: http://dx.doi.org/10.1016/S1473-3099(04)01253-8
Fewtrell, L., Prüss-Üstün, A., Bos, R., Gore, F., & Bartram, J. (2007). Water, Sanitation and Hygiene Quantifying the Health Impact at National Local Levels in Countries with Incomplete Waters Supply and Sanitation Coverage. Environmental Burden of Disease Series (Vol. 15. ). Geneva, Switzerland: World Health Organization.
FMoH. (2005). National hygiene and sanitation strategy To enable 100% adoption of improved hygiene and sanitation. Addis Ababa, Ethiopia: Federal Democratic Republic of Ethiopia Ministry of Health (FMoH).
FMoH. (2010). Health sector development programme III Annual performance report EFY 2002 (2009/2010) (2 ed.). Addis Ababa, Ethiopia: The Federal Democratic Republic of Ethiopia.
FMoH. (2011). National hygiene & sanitation strategic action plan for rural, per-urban and informal settlements in Ethiopia (Vol. Part II). Addis Ababa, Ethiopia: Federal Democratic Republic of Ethiopia Ministry of Health (FMoH).
138
Gebregziabher, G. (2008). Risk and irrigation investment in a semi-arid economy. (PhD) Thesis, Department of Economics and Resource Management, Norwegian University of Life Sciences, Norway, Aos.
Ghebreyesus, T. A., Haile, M., Witten, K. H., Getachew, A., Yohannes, A. M., Yohannes, M., . . . Byass, P. (1999). Incidence of malaria among children living near dams in northern Ethiopia: community based incidence survey. BMJ, 319(7211), 663-666. doi: 10.1136/bmj.319.7211.663
Gilman, R. H., Marquis, G. S., Ventura, G., Campos, M., Spira, W., & Diaz, F. (1993). Water cost and availability: key determinants of family hygiene in a Peruvian shantytown. American Journal of Public Health, 83(11), 1554-1558.
Glewwe, P., Koch, S., & Nguyen, B. L. (2002). Child nutrition, economic growth, and the provision of health care services in Vietnam in the 1990s (Vol. 2776): World Bank, Development Research Group, Macroeconomics and Growth.
Godfrey, S., Labhasetwar, P., Wate, S., & Pimpalkar, S. (2011). How safe are the global water coverage figures? Case study from Madhya Pradesh, India. Environmental Monitoring and Assessment, 176(1-4), 561-574. doi: 10.1007/s10661-010-1604-3
Graham, J. P., & Polizzotto, M. L. (2013). Pit latrines and their impacts on groundwater quality: A systematic review. Environmental health perspectives, 121.
Gray, V. B., Cossman, J. S., & Powers, E. L. (2006). Stunted growth is associated with physical indicators of malnutrition but not food insecurity among rural school children in Honduras. Nutrition Research, 26(11), 549-555. doi: 10.1016/j.nutres.2006.09.009
Greene, W. H. (2012). Econometric Analysis (Seventh ed.). New York University, USA: Prentice Hall, Pearson Education International.
Guerrant, R. L., Oriá, R. B., Moore, S. R., Oriá, M. O. B., & Lima, A. A. M. (2008). Malnutrition as an enteric infectious disease with long-term effects on child development. Nutrition Reviews, 66(9), 487-505. doi: 10.1111/j.1753-4887.2008.00082.x
Gundry, S., Conroy, R., & Wright, J. (2004). A systematic review of the health outcomes related to household water quality in developing countries. J. Water Health, 2(1), 1-13.
Günther, I., & Fink, G. (2010). Water, Sanitation and Children's Health. Evidence from 172 DHS Surveys. World Bank Policy Research Working Paper No. 5275. .
Günther, I., & Schipper, Y. (2013). Pumps, germs and storage: The impact of improved water containers on water quality and health. Health Economics, 22(7), 757-774. doi: 10.1002/hec.2852
Haggerty, P. A., Muladi, K., Kirkwood, B. R., Ashworth, A., & Manunebo, M. (1994). Community-based hygiene education to reduce diarrhoeal disease in rural Zaire: impact of the intervention on diarrhoeal morbidity. International journal of epidemiology, 23(5), 1050-1059.
Hagos, F., Boelee, E., Awulachew, S. B., Slaymaker, T., Tucker, J., & Ludi, E. (2008). Water supply and sanitation (WSS) and Poverty, Micro-level linkages in Ethiopia. (RiPPLE Working Papers Issue 8).
Hawkes, C., & Ruel, M. (2006). The Links Between Agriculture and Health: An Intersectoral Opportunity to Improve the Health and Livelihoods of the Poor. Bulletin of the World Health Organization, 84(12), 984-990.
139
Haylamicheal, D., & Moges, A. (2012). Assessing water quality of rural water supply schemes as a measure of service delivery sustainability: A case study of WondoGenet district, Southern Ethiopia. African Journal of Environmental Science and Technology, 6(5), 229-236.
Headey, D. (2014). An analysis of trends and determinants of child undernutrition in Ethiopia, 2000-2011 (Vol. Working paper 70): International Food Policy Research Institute (IFPRI).
Hebert, J. R. (1985). Effects of water quality and water quantity on nutritional status: findings from a south Indian community. Bulletin of the World Health Organization, 63(1), 143.
Hill, Z., Kirkwood, B., & Edmond, K. (2004). Family and community practices that promote child survival, growth and development: a review of the evidence
Horgby, Å., & Larson, K. (2013). Effects on Drinking Water Quality Due to Irrigation in the Koga area of Merawi, Ethiopia.
Howard, G., & Bartram, J. (2003). Domestic water quantity, service level, and health. Water Engineering and Development Centre (WEDC), Loughborough University, UK, and Water, Sanitation and Health Programme (WHO/SDE/WSH/03.02). Retrieved from http://www.who.int/water_sanitation_health/diseases/WSH0302.pdf
Humphrey, J. H. (2009). Child undernutrition, tropical enteropathy, toilets, and handwashing. Lancet (London, England), 374(9694), 1032-1035.
Hutton, G., & Haller, L. (2004). Evaluation of the costs and benefits of water and sanitation improvements at the global level: Water, Sanitation, and Health, Protection of the Human Environment, World Health Organization.
IFPRI. (2015). 2014-2015 Global food policy report. Washington, D.C.: International Food Policy Research Institute (IFPRI). http://dx.doi.org/10.2499/9780896295759.
ILRI. (2005). Fogera woreda pilot learning Site diagnosis and program design and atlas. Nairobi, Kenya: International Livestock Research Institute (ILRI).
IRC. (2007). Towards Effective Programming for WASH in Schools: A Mannual on Scaling up Programmes for Water, Sanitation and Hygiene in Schools. Delft, The Netherlands: IRC International Water and Sanitation Center (TP series; no. 48). 93p.
Jalan, J., & Ravallion, M. (2003). Does Piped Water Reduce Diarrhea for Children in Rural India? Journal of Econometrics, 112(1), 153-173. doi: http://dx.doi.org/10.1016/S0304-4076(02)00158-6
Jano, M. (2007). Assessment of Physicochemical and microbiological Quality of rural Drinking water supply at the source in Menge Woreda, Benshangul Gumuz Regional state, Ethiopia. M. Sc Thesis. Addis Ababa University, Addis Ababa.
Kabubo-Mariara, J., Ndenge, G. K., & Mwabu, D. K. (2009). Determinants of children's nutritional status in Kenya: Evidence from demographic and health surveys. Journal of African Economies, 18(3), 363-387.
Keiser, J., De Castro, M. C., Maltese, M. F., Bos, R., Tanner, M., Singer, B. H., & Utzinger, J. (2005). Effect of irrigation and large dams on the burden of malaria on a global and regional scale. American Journal of Tropical Medicine and Hygiene, 72(4), 392-406.
Kennedy, E. T., & Bouis, H. E. (1993). Linkages Between Agriculture and Nutrition: Implications for Policy and Research: International Food Policy Research Institute (IFPRI).
140
Kibret, S., Alemu, Y., Boelee, E., Tekie, H., Alemu, D., & Petros, B. (2010). The impact of a small-scale irrigation scheme on malaria transmission in Ziway area, Central Ethiopia. Tropical medicine & international health, 15(1), 41-50. doi: 10.1111/j.1365-3156.2009.02423.x
Kibret, S., McCartney, M., Lautze, J., & Jayasinghe, G. (2009). Malaria transmission in the vicinity of impounded water: Evidence from the Koka Reservoir, Ethiopia (Vol. 132). Colombo, Sri Lanka: International Water Management Institute (IWMI). 39p. (IWMI Research Report 132) doi: 10.3910/2009.129: Iwmi.
Kinfegabriel, T. (2014). Welthungerhilfe Ethiopia WASH Sector Experience. A presentation at WATSAN Agriculture Workshop Organized by Center for Development Research (ZEF) of University of Bonn, Germany, 8-9 September.
Kirchberger, M. (2008). Do Households’ Water and Sanitation Choices Really Matter for Child Health? Master thesis. University of Oxford.
Kirogo, V., Wambui, K., & Muroki, N. (2007). The role of irrigation on improvement of nutritional status of young children in central Kenya. African Journal of Food, Agricult Nutrition and Development, 7(2), 1-16.
Komatsu, H., Malapit, H. J. L., & Theis, S. (2015). How does women’s time in reproductive work and agriculture affect maternal and child nutrition? Evidence from Bangladesh, Cambodia, Ghana, Mozambique, and Nepal. . IFPRI Discussion Paper 1486. Washington, D.C.: International Food Policy Research Institute (IFPRI). http://ebrary.ifpri.org/cdm/ref/collection/p15738coll2/id/129842.
Koolwal, G., & Van de Walle, D. (2010). Access to Water, Women's Work and Child Outcomes. World Bank Policy Research Working Paper No.5302.
Koolwal, G., & Walle, D. v. d. (2013). Access to Water, Women’s Work, and Child Outcomes. Economic Development and Cultural Change, 61(2), 369-405. doi: 10.1086/668280
Kremer, M., Leino, J., Miguel, E., & Zwane, A. P. (2009). Spring cleaning: rural water impacts, valuation and property rights institutions: National Bureau of Economic Research.
Kremer, M., & Zwane, A. P. (2007). Cost-effective prevention of diarrheal diseases: A critical review. Center for Global Development Working Paper(117).
Kumar, S., & Vollmer, S. (2013). Does Access to Improved Sanitation Reduce Childhood Diarrhea in Rural India? Health Economics, 22(4), 410-427. doi: 10.1002/hec.2809
Kumie, A., & Ali, A. (2005). An overview of environmental health status in Ethiopia with particular emphasis to its organization, drinking water and sanitation: a literature survey. Ethiopian Journal of Health Development, 19(2), 89.
Larson, B., Minten, B., & Razafindralambo, R. (2006). Unravelling the linkages between the millennium development goals for poverty, education, access to water and household water use in developing countries: Evidence from Madagascar. The Journal of Development Studies, 42(1), 22-40. doi: 10.1080/00220380500356258
Levy, K., Nelson, K. L., Hubbard, A., & Eisenberg, J. N. (2008). Following the water: a controlled study of drinking water storage in northern coastal Ecuador. Environ Health Perspect, 116(11), 1533-1540.
Lin, A., Arnold, B. F., Afreen, S., Goto, R., Huda, T. M. N., Haque, R., . . . Luby, S. (2013). Household Environmental Conditions Are Associated with Enteropathy and Impaired Growth in Rural Bangladesh. The American journal of tropical medicine and hygiene, 89(1), 130-137. doi: 10.4269/ajtmh.12-0629
141
Li, V. (2014). The rise, critique and persistence of the DALY in global health. The Journal of Global Health.
Lorntz, B., Soares, A. M., Moore, S. R., Pinkerton, R., Gansneder, B., Bovbjerg, V. E., . . . Guerrant, R. L. (2006). Early Childhood Diarrhea Predicts Impaired School Performance. The Pediatric Infectious Disease Journal, 25(6), 513-520 510.1097/1001.inf.0000219524.0000264448.0000219590.
Luby, S., Agboatwalla, M., Feikin, D. R., Painter, J., Billhimer, W., Altaf, A., & Hoekstra, R. M. (2005). Effect of handwashing on child health: a randomised controlled trial. Lancet, 366(9481), 225-233.
Luby, S., Agboatwalla, M., Painter, J., Altaf, A., Billhimer, W., Keswick, B., & Hoekstra, R. M. (2006). Combining drinking water treatment and hand washing for diarrhoea prevention, a cluster randomised controlled trial. Tropical medicine & international health, 11(4), 479-489. doi: 10.1111/j.1365-3156.2006.01592.x
MacKinnon, J. G., & Magee, L. (1990). Transforming the dependent variable in regression models. International Economic Review, 315-339.
Mangyo, E. (2008). The effect of water accessibility on child health in China. Journal of Health Economics, 27(5), 1343-1356. doi: http://dx.doi.org/10.1016/j.jhealeco.2008.04.004
Mara, D., Lane, J., Scott, B., & Trouba, D. (2010). Sanitation and Health. PLoS Med, 7(11), e1000363. doi: 10.1371/journal.pmed.1000363
Megha, P. U., Kavya, P., Murugan, S., & Harikumar, P. S. (2015). Sanitation Mapping of Groundwater Contamination in a Rural Village of India. Journal of Environmental Protection, Vol.06No.01, 11. doi: 10.4236/jep.2015.61005
Mengistie, B., Berhane, Y., & Worku, A. (2013). Household Water Chlorination Reduces Incidence of Diarrhea among Under-Five Children in Rural Ethiopia: A Cluster Randomized Controlled Trial. PLoS ONE, 8(10), e77887. doi: 10.1371/journal.pone.0077887
Miller, G., & Urdinola, B. P. (2010). Cyclicality, Mortality, and the Value of Time: The Case of Coffee Price Fluctuations and Child Survival in Colombia. The Journal of Political Economy, 118(1), 113-155.
Mintz, E. D., Reiff, F. M., & Tauxe, R. V. (1995). Safe water treatment and storage in the home: a practical new strategy to prevent waterborne disease. The Journal of the American Medical Association, 273(12), 948-953. doi: doi:10.1001/jama.1995.03520360062040
MoARD. (2009). Small scale irrigation capacity building strategy for Ethiopia: Output of a tripartite cooperation between the Federal Republic of Germany, the state of Israel and the Federal Democratic Republic of Ethiopia. Addis Ababa, Ethiopia: Ministry of Agriculture and Rural Development (MoARD).
Morris, S. S., Cousens, S. N., Kirkwood, B. R., Arthur, P., & Ross, D. A. (1996). Is Prevalence of Diarrhea a Better Predictor of Subsequent Mortality and Weight Gain Than Diarrhea Incidence? American Journal of Epidemiology, 144(6), 582-588.
MoWE. (2006). Universal Access Program for water Supply and Sanitatio Addis Ababa, Ethiopia: Ministry of Water and Energy.
MoWE. (2007). Water sector development programme (WSDP) reports. Addis Ababa, Ethiopia: Ministry of Water and Energy (MoWE).
Mutero, C. M., McCartney, M., & Boelee, E. (2006). Agriculture, Malaria, and Water-associated Diseases. In C. Hawkes & M. T. Ruel (Eds.), Understanding the links Between Agriculture and Health. Washington, D.C:: International Food Policy research institute (IFPRI).
142
Mwabu, G. (2007). Health Economics for Low-Income Countries. In T. P. Schultz & A. S. John (Eds.), Handbook of Development Economics (Vol. 4, pp. 3305-3374): Elsevier.
Namara, R., Upadhyay, B., & Nagar, R. K. (2005). Adoption and impacts of microirrigation technologies: Empirical results from selected localities of Maharashtra and Gujarat States of India (Vol. 93): IWMI.
Namara, R., Awuni, J. A., Barry, B., Giordano, M., Hope, L., Owusu, E. S., & Forkuor, G. (2011). Smallholder shallow groundwater irrigation development in the upper east region of Ghana (Vol. 143): IWMI.
Nichols, A. (2011). Causal inference for binary regression. Paper presented at the Stata conference Chicago,(version June.
Olney, D. K., Talukder, A., Iannotti, L. L., Ruel, M. T., & Quinn, V. (2009). Assessing Impact and Impact Pathways of a Homestead Food Production Program on Household and Child Nutrition in Cambodia. Food and Nutrition Bulletin, 30(4), 355-369. doi: 10.1177/156482650903000407
Osman, M., & Sauerborn, P. (2002). Rainfall variability and its influence on surface flow regimes examples from the central highlands of Ethiopia. Journal of Soils and Sediments, 2(2), 100-102.
Overbey, L. (2008). The Health Benefits of Water Supply and Sanitation Projects: A Review of the World Bank Lending Portfolio IEG Working Paper 2008/1.
Patwari, A. K. (1999). Diarrhoea and malnutrition interaction. Indian journal of pediatrics, 66(1 Suppl), S124-134.
Pickering, A. J., & Davis, J. (2012). Freshwater Availability and Water Fetching Distance Affect Child Health in Sub-Saharan Africa. Environmental science & technology, 46(4), 2391-2397. doi: 10.1021/es203177v
Prüss-Üstün, A., Bos, R., Gore, F., & Bartram, J. (2008). Safer Water, Better Health: Costs, Benefits and Sustainability of Interventions to Protect and Promote Health: World Health Organization.
Prüss-Üstün, A., & Corvalán, C. (2006). Preventing disease through healthy environments. Towards an Estimate of the Environmental Burden of Disease: World Health Organization, Geneva.
Prüss-Üstün, A., Kay, D., Fewtrell, L., & Bartram, J. (2002). Estimating the burden of disease from water, sanitation, and hygiene at a global level. Environmental Health Perspectives, 110(5), 537-542. Retrieved from http://www.who.int/quantifying_ehimpacts/global/en/ArticleEHP052002.pdf.
Ray, I. (2007). Women, Water, and Development. Annual Review of Environment and Resources, 32(1), 421-449. doi: doi:10.1146/annurev.energy.32.041806.143704
Reed, B. (2005). Minimum water quantity needed for domestic use in emergencies. WHO – Technical Notes for Emergencies.
Rogers, P., Bhatia, R., & Huber, A. (1998). Water as a social and economic good: How to put the principle into practice. Global Water Partnership Technical Advisory Committee Background Paper No. 2.
Roushdy, R., Sieverding, M., & Radwan, H. (2012). The Impact of Water Supply and Sanitation on Child Health: Evidence from Egypt. Population Council.
Rufener, S., Mäusezahl, D., Mosler, H.-J., & Weingartner, R. (2010). Quality of Drinking-water at Source and Point-of-consumption—Drinking Cup As a High Potential Recontamination Risk: A Field Study in Bolivia. Journal of Health, Population, and Nutrition, 28(1), 34-41.
143
Sandiford, P., Gorter, A. C., Davey Smith, G., & Pauw, J. P. C. (1989). Determinants of drinking water quality in rural Nicaragua. Epidemiology & Infection, 102(03), 429-438. doi: doi:10.1017/S0950268800030144
Save the Children. (2012). A life Free From Hunger: Tackling Child Malnutrition. London, UK: Save the Children.
Scheelbeek, P. (2005). Two Containers a Day: The Search for proper Water Sources in Eastern Ethiopia. Water Quality and Sanitation in the Lege Dini Watershed Area, Ethiopia. MSc Thesis in Irrigation and Water Engineering. Wageningen University, the Netherlands.
Schorling, J. B., McAuliffe, J. F., de Souza, M. A., & Guerrant, R. L. (1990). Malnutrition is associated with increased diarrhoea incidence and duration among children in an urban Brazilian slum. International Journal of Epidemiology, 19(3), 728-735.
Sorenson, S. B., Morssink, C., & Campos, P. A. (2011). Safe access to safe water in low income countries: Water fetching in current times. Social Science & Medicine, 72(9), 1522-1526. doi: http://dx.doi.org/10.1016/j.socscimed.2011.03.010
Spears, D. (2013). How much International Variation in Child Height Can Sanitation Explain? World Bank Policy Research Working Paper 6351, Available at: http://www-wds.worldbank.org/external/default/WDSContentServer/IW3P/IB/2013/02/05/000158349_20130205082533/Rendered/PDF/wps6351.pdf.
Stel, N., & Abate, F. M. (2014). Between Control and Cooperation: Multi-Stakeholder Service Provision and the Legitimacy of State Institutions in Ethiopia’s Amhara National Regional State. European Journal of Development Research, 26(5), 743-760.
Sutton, S., Butterworth, J., & Mekonta, L. (2012). A hidden resource: household-led rural water supply in Ethiopia: IRC International Water and Sanitation Centre, The Netherlands.
Sutton, S., Mamo, E., Butterworth, J., & Dimtse, D. (2011). Towards the Ethiopian goal of universal access to rural water: understanding the potential contribution of self supply: RiPPLE Working Paper.
Thomas, D., Strauss, J., & Henriques, M.-H. (1991). How Does Mother's Education Affect Child Height? The Journal of Human Resources, 26(2), 183-211. doi: 10.2307/145920
Tilahun, S. A., Tigabu, A. D., Tarekegne, T. M., Addisie, M. B., Beyene, H. A., Alemeyehu, Z. A., . . . Steenhuis, T. S. (2013). Factors in the suboptimum performance of rural water supply systems in the Ethiopian highlands. Retrieved from http://hdl.handle.net/10568/34247 (Accessed 15 March 2015)
Trevett, A. F., Carter, R. C., & Tyrrel, S. F. (2005). Mechanisms leading to post - supply water quality deterioration in rural Honduran communities. International Journal of Hygiene and Environmental Health, 208(3), 153-161. doi: http://dx.doi.org/10.1016/j.ijheh.2005.01.024
Tsega, N., Sahile, S., Kibret, M., & Abera, B. (2014). Bacteriological and physico-chemical quality of drinking water sources in a rural community of Ethiopia. African health sciences, 13(4), 1156-1161.
Tsegai, D. W., McBain, F., & Tischbein, B. (2013). Water, sanitation and hygiene: The missing link with agriculture. (ZEF Working Paper Series, No.107).
Ulijaszek, S. J. (1996). Relationships between undernutrition, infection, and growth and development. Human Evolution, 11(3-4), 233-248. doi: 10.1007/bf02436627
Resolution Adopted by the Agency Assembly: The Human Right to Water and Sanitation, A/RES/64/292 C.F.R. (2010).
144
UN Water. (2008a). Tackling a Global Crises: International Year of Sanitation 2008: Available at: http://esa.un.org/iys/docs/IYS_flagship_web_small.pdf.
UN Water. (2008b). Sanitation: A Wise Investment for Health, Dignity, and Development. Key Messages for the International Year of Sanitation, 2008. Available at: http://www.preventionweb.net/files/2378_IYS20keymessage20booklet.pdf.
UNDP. (2006). Human Development Report 2006-Beyond scarcity: Power, poverty and the global water crisis. New York: USA: Palgrave Macmillian.
UNICEF. (2000). Sanitation for all: Promoting dignity and human rights. UNICEF, New York.
UNICEF. (2009). State of the World's Children 2009: Maternal and Newborn Health New York, USA: United Nations Children's Fund (UNICEF).
UNICEF. (2012). Committing to Child Survival: A Promise Renewed, Progress Report 2012. New York: United Nations Children's Fund (UNIEF).
UNICEF. (2013). Improving child nutrition: The achievable imperatives for global progress. New York: United Nations Children’s Fund (UNICEF). http://www.childinfo.org/files/NutritionReport_April2013_Final.pdf.
Usman, M. A., Gerber, N. J. von Braun (2016). The Impact of Drinking Water Quality and Sanitation on Child Health: Evidence from Rural Ethiopia. Discussion Paper No. 221. Center for Development Research (ZEF). Bonn, Germany.
Usman, M. A., Gerber, N., & Pangaribowo, E. H. (2016). Determinants of Household Drinking Water Quality in Rural Ethiopia. Discussion Paper No. 220. Center for Development Research (ZEF). Bonn, Germany.
van der Hoek, W., Feenstra, S. G., & Konradsen, F. (2002). Availability of irrigation water for domestic use in Pakistan: its impact on prevalence of Diarrhoea and Nutritional Status of Children. Journal of Health, Population and Nutrition, 20(1), 77-84.
van Der Hoek, W., Konradsen, F., Ensink, J. H. J., Mudasser, M., & Jensen, P. K. (2001). Irrigation water as a source of drinking water: is safe use possible? Tropical medicine & international health, 6(1), 46-54. doi: 10.1046/j.1365-3156.2001.00671.x
van Der Hoek, W., Konradsen, F., & Jehangir, W. A. (1999). Domestic Use of Irrigation Water: Health Hazard or Opportunity? International Journal of Water Resources Development, 15(1-2), 107-119. doi: 10.1080/07900629948961
van Koppen, B., Smits, S., & Mikhail, M. (2009). Homestead- and community-scale multiple-use water services: unlocking new investment opportunities to achieve the Millennium development Goals. Irrigation and Drainage, 58(S1), S73-S86. doi: 10.1002/ird.486
van Koppen, B., Smits, S., Moriarty, P., Penning de vries, F., Mikhail, M., & Boelee, E. (2009). Climbing the water ladder: Multiple use water services for poverty reduction. The Hague, the Netherlands: IRC International Water and Sanitation Center and International Water Management Institute (TP, series, no. 52) 213p.
Viscusi, W. K. & Aldy, J. E. (2002). The value of a statistical life: A critical review of market estimates throughout the world. Harvard Law School John M. Olin Center for Law, Economics and Business Discussion Paper Series 392. Retrieved from http://lsr.nellco.org/harvard_olin/392.
145
von Braun, J., Puetz, D., & Webb, P. (1989). Irrigation technology and commercialization of rice in the Gambia: Effects on income and nutrition (Vol. IFPRI Research Report 75). Washington, DC: International Food Policy Research Institute (IFPRI).
von Braun, J., Ruel, M. T., & Gillespie, S. (2012). Bridging the Gap between the Agriculture and Health Sectors. In Shenggen Fan & R. Pandya-Lorch (Eds.), Reshaping Agriculture for Better Nutrition (pp. 183). Washington, DC.: International Food Policy Research Institute (IFPRI).
Waddington, H., & Snilstveit, B. (2009). Effectiveness and sustainability of water, sanitation, and hygiene interventions in combating diarrhoea. Journal of Development Effectiveness, 1(3), 295-335. doi: 10.1080/19439340903141175
Waddington, H., Snilstveit, B., White, H., & Fewtrell, L. (2009). Water, sanitation and hygiene interventions to combat childhood diarrhoea in developing countries (Vol. 31): International Initiative for Impact Evaluation New Delhi.
Wagner, E. G., & Lanoix, J. N. (1959). Water supply for rural areas and small communities.
Walker, C. L.F., Perin, J., Aryee, M. J., Boschi-Pinto, C., & Black, R. E. (2012). Diarrhea incidence in low- and middle-income countries in 1990 and 2010: a systematic review. [journal article]. BMC Public Health, 12(1), 1-7. doi: 10.1186/1471-2458-12-220
Wang, X., & Hunter, P. R. (2010). Short Report: A systematic review and meta-analysis of the association between self-reported diarrheal disease and distance from home to water source. The American journal of tropical medicine and hygiene, 83(3), 582-584.
WELL. (1998). Guidance Manual on Water Supply and Sanitation Programmes: WEDC, Loughborough University.
Welle, K., & Williams, J. (2014). Monitoring and addressing governance factors affecting rural water supply sustainability. Regional approach paper based on Global Water Initiative East Africa Governance-into-Functionality monitoring of water schemes implemented in Ethiopia, Tanzania and Uganda. GWI East Africa.
Whinnery, J. (2012). A Well Construction Cost-Benefit Analysis (CBA): For Water Supply Well Guidelines for use in Developing Countries.
WHO. (1996). Water and Sanitation. Geneva, Switzerland: World Health Organization; Fact Sheet Number 112.
WHO. (2002). The World Health Report 2002: Reducing Risks, Promoting Health Life. Geneva, Switzerland.
WHO. (2011). Guideline for Drinking-water Quality. Geneva, Switzerland: World Health Organization
WHO Multicentre Growth Reference Study Group. (2006). WHO child growth standards: Length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age: Methods and development. Geneva: World Health Organization.
WHO. (2012). Burden of disease and cost-effectiveness estimates. Retrieved from http://www.who.int/water_sanitation_health/diseases/burden/ en/index.html.
WHO/UNICEF. (2009). Diarrhoea: Why children are Still Dying and what can be done: World Health Organization (WHO) and United Nations Children's Fund (UNICEF).
146
WHO/UNICEF. (2010a). Rapid Assessment of Drinking Water Quality in the Federal Republic of Ethiopia: Country Report, World Health Organization (WHO), Federal Ministry of Health (FMoH), Ethiopian Health & Nutrition Research Institute (EHNRI) and United Nations Children's Fund (UNICEF).
WHO/UNICEF. (2010b). Progress on Sanitation and Drinking-Water, 2010 Update. Available at: http://whqlibdoc.who.int/publications/2010/9789241563956_eng_full_text.pdf.
WHO/UNICEF. (2012). Progress on Drinking Water and Sanitation: 2012 Update. New York, NY, USA, United Nations Children’s Fund (UNICEF); Geneva, Switzerland, World Health Organization (WHO)
WHO/UNICEF. (2014). Progress on Drinking Water and Sanitation: 2014 Update. New York, NY, USA, United Nations Children’s Fund (UNICEF); Geneva, Switzerland, World Health Organization (WHO)
WHO/UNICEF. (2015). Progress on Drinking Water and Sanitation: 2014 Update and MDG Assessment. New York, NY, USA, United Nations Children’s Fund (UNICEF); Geneva, Switzerland, World Health Organization (WHO)
Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data: MIT press.
World Bank. (2005). Ethiopia: A country status report on health and poverty: Africa Region Human Development, The World Bank, Ministry of Health, Ethiopia.
Wright, J., Gundry, S., & Conroy, R. (2004). Household drinking water in developing countries: a systematic review of microbiological contamination between source and point-of-use. Tropical medicine & international health, 9(1), 106-117. doi: 10.1046/j.1365-3156.2003.01160.x
Yasin, M., Ketema, T., & Bacha, K. (2015). Physico-chemical and bacteriological quality of drinking water of different sources, Jimma zone, Southwest Ethiopia. BMC research notes, 8(1), 541.
Zwane, A. P., & Kremer, M. (2007). What Works in Fighting Diarrheal Diseases in Developing Countries? A Critical Review. The World Bank Research Observer, 22(1), 1-24. doi: 10.1093/wbro/lkm002
147
Appendix A
Table A1: Impact of latrine on household water quality: Number of E.coli (CFU/100ml water)
(1) (2) (3)
VARIABLES OLS SE OLS SE PROBIT SE
Pit latrine (dummy) 23.088*** 6.597 15.368** 6.467 0.098 0.100
Drinking water source (1=protected) -15.885** 7.048 -8.711 6.852 0.185 0.125
Container (1=jerrycan) 19.255** 8.428 0.675** 0.323
Irrigating households (dummy) -0.796 8.107 0.885** 0.374
Handwashing with soap(dummy) -0.044 7.207 0.090 0.088
Number of adult female -7.078 7.699 -0.052 0.100
Household density 2.851 2.648 0.120* 0.063
Household size -1.274 2.769 0.006 0.033
Number of children aged 6/younger 8.539 5.357 -0.430** 0.186
Garbage disposal (reference burning)
Throw in the yard 17.857** 8.458 0.559*** 0.062
Throw away outside the yard -2.414 10.019 -0.222*** 0.079
Used as fertilizer 6.601 9.586 0.416*** 0.073
Livestock holding 5.407* 2.704 0.052 0.035
Log of assets value -5.720 4.459 -0.393** 0.191
Constant 16.187*** 5.105 16.169 38.671
Observations 71 71 71
Pseudo/R-squared 0.20 0.51 0.75
Model F-stat 8.363 4.232
Model Chi2 71.88
Model p-value 0.000 0.000 0.000
Note: the dependent variable for column 1 and 2 is the log of E.coli while for column 3 is binary. Significance *** p<0.01, ** p<0.05, * p<0.1
148
Table A2: Impact of irrigation and domestic use of irrigation water on p/c/d water consumption
(1) (2) (3) (4) (5) VARIABLES Model Model Model Model Model
Irrigating households (dummy) 0.577** 0.391 0.436** (0.274) (0.252) (0.214) Domestic use of irrigation water (dummy) 0.425 0.735** (0.313) (0.277) Container size in liters 0.208*** 0.248*** 0.203*** 0.247*** (0.0397) (0.033) (0.051) (0.043) Minutes to water source (1=30min/less) 1.232*** 1.065*** 1.343*** 1.327*** (0.228) (0.186) (0.368) (0.327) Proportion of adult women 9.109*** 2.277* 9.386*** 2.112 (1.593) (1.293) (1.979) (1.549) Household size -0.686*** -0.731*** (0.059) (0.070) Log of assets value 0.213 0.323** (0.135) (0.138) District (1=Fogera) 0.545*** 0.815*** (0.189) (0.269) Constant 8.387*** 0.979 3.232*** 1.168 2.233* (0.219) (1.001) (1.170) (1.245) (1.254)
Observations 454 454 454 302 302 R-squared 0.01 0.19 0.38 0.19 0.40 Model F-stat 4.418 28.45 43.08 15.86 30.09 Model p-value 0.0398 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering at village level in parenthesis; Significance *** p<0.01, ** p<0.05, * p<0.1
149
Table A3: The impact of domestic use of irrigation water on POU water quality
(1) (2) VARIABLES OLS SE PROBIT SE
Irrigation water use (dummy) -0.110 0.209 -0.092 0.058 Water source (1=protected) -0.202 0.150 -0.074 0.052 Minutes to water source (1=30min/less) -0.890*** 0.293 -0.176* 0.097
Container (1=jerrycan) 1.151*** 0.162 0.231*** 0.066 Number of adult women -0.192 0.144 -0.076* 0.044
Household density 0.297*** 0.060 0.072*** 0.018 Garbage disposal (reference burning) Throw in the yard 1.107*** 0.213 0.346*** 0.074 Throw away outside the yard 0.205 0.359 0.032 0.099
Used as fertilizer 0.730** 0.285 0.208** 0.095 Handwashing with soap (dummy) -0.326** 0.152 -0.104** 0.045 Highest education completed -0.052 0.032 -0.015** 0.006 Livestock holding 0.273*** 0.059 0.066*** 0.017 Log of assets value -0.321*** 0.081 -0.051** 0.020 Water user association(dummy) -1.123*** 0.187 -0.354*** 0.063 Distance to irrigation water (hours) -0.686** 0.274 -0.237*** 0.090
Constant 3.024*** 0.808
Observations 302 302
Pseudo / R-squared 0.48 0.41 Model F-stat 46.30
Model Ch2 142.66 Model p-value 0.000 0.000
Robust standard errors adjusted for clustering at the village level; Significance *** p<0.01, ** p<0.05, * p<0.1 Probit in average marginal effects Note: the dependent variable for column 1 is log of E.coli while for column 2 is binary.
150
Table A4: The effect of irrigation on primary caretaker time use
(1) (2) VARIABLES Water fetching SE Agriculture SE
Age in years -0.306 0.198 -0.135 0.159 Irrigating households (dummy) -13.565*** 3.318 4.552** 1.828 Household size -2.135** 0.924 -1.278* 0.685
Proportion of adult women -43.337** 18.674 Number of children age 6/younger 2.303 2.546 2.440* 1.230
Water collection time (1=30min/less) -16.123*** 4.073 2.277 1.851 Container size -0.558 0.501 Log of off-farm income 1.128*** 0.418 0.207 0.196 Own land holding -3.644** 1.639 1.135 1.098 Livestock holding 1.270* 0.652 0.471 0.444
Proportion of irrigated land 5.840** 2.900 Constant 86.177*** 14.818 18.176*** 4.434
Observations 453 453
R-squared 0.19 0.09 Model F-stat 10.98 5.026 Model p-value 0.000 0.000
Robust standard errors adjusted for clustering at village level; Significance *** p<0.01, ** p<0.05, * p<0.1 Note: Water fetching time is reported in minutes for the day before the interview day while agricultural activity is reported in the last 7 days before the survey and reported in hours.
151
Table A5: Health effects of irrigation—Water-related and morbidity
Water related Morbidity
VARIABLES (1) (2) (3) (4) (5) (6)
Irrigating households (dummy) 0.039** 0.110* 1.458** 0.110*** -0.023 1.448***
(0.018) (0.060) (0.245) (0.031) (0.050) (0.160)
Water quality (1= no E.coli) -0.011 -0.024 0.920 -0.043* -0.060 0.880*
(0.017) (0.054) (0.138) (0.022) (0.042) (0.065)
Water collection time (1=30min/less) -0.001 -0.051 0.975 -0.006 -0.046 0.994
(0.017) (0.067) (0.149) (0.026) (0.052) (0.088)
Pit latrine (dummy) -0.032** -0.136*** 0.751** 0.039* -0.031 1.145*
(0.014) (0.044) (0.104) (0.022) (0.037) (0.083)
Age in years: reference (5—6)
7—9 years -0.002 -0.200***
(0.021) (0.048)
10—15 years 0.034 -0.255***
(0.023) (0.047)
16 years/older 0.060*** -0.171***
(0.020) (0.044)
Gender (male=1) -0.003 -0.037
Average household member age -0.000 1.000 -0.002 0.976**
(0.006) (0.013) (0.004) (0.011)
Household size -0.008 0.047** 1.107** 0.005 0.061*** 1.209***
(0.005) (0.018) (0.048) (0.008) (0.014) (0.033)
Household density 0.013** 0.026 1.118** 0.017* 0.045*** 1.045
(0.006) (0.019) (0.050) (0.009) (0.016) (0.030)
Highest education completed -0.002 -0.006 0.993 -0.011*** -0.018*** 0.966***
(0.002) (0.008) (0.022) (0.003) (0.006) (0.012)
Distance to health center 0.001 0.005 1.007 0.009*** 0.011** 1.027***
(0.002) (0.004) (0.014) (0.003) (0.004) (0.009)
Livestock holding 0.004 0.006 1.052 0.004 0.007 1.032
(0.005) (0.020) (0.052) (0.009) (0.015) (0.029)
Log of assets value 0.012 0.034 1.131 0.008 0.038 1.035
(0.010) (0.036) (0.104) (0.015) (0.028) (0.049)
Proportion of irrigated land -0.009 -0.073 0.927 -0.105** -0.044 0.700***
(0.026) (0.085) (0.204) (0.041) (0.062) (0.090)
Constant 0.053*** 0.370**
(0.040) (0.176)
Observation 2149 454 454 2149 454 454
Pseudo R-squared 0.03 0.07 0.05 0.15
Model Chi2 54.00 46.16 59.62 147.10 74.06 266.74
Model p-value 0.000 0.000 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1; Probit in average marginal effects and Poisson in incidence-rate ratios.
152
Table A6: Health effects of irrigation water use—Water related
(1) (2) (3) (4) (5) (6)
VARIABLES PROBIT PROBIT PROBIT PROBIT POISSON POISSON
Irrigation water use (dummy) 0.046*** 0.046*** 0.127** 0.126** 1.475*** 1.476***
(0.017) (0.016) (0.052) (0.051) (0.205) (0.199)
Water quality (1= no E.coli) -0.002 0.014 0.992
(0.023) (0.069) (0.179)
Water collection time (1=30min/less) -0.010 -0.010 -0.149 -0.146 0.927 0.926
(0.022) (0.021) (0.096) (0.093) (0.149) (0.144)
Pit latrine (dummy) -0.021 -0.021 -0.093 -0.093 0.854 0.855
(0.019) (0.019) (0.066) (0.066) (0.142) (0.142)
Age in years: reference (5—6)
7—9 years 0.005 0.005
(0.027) (0.027)
10—15 years 0.046 0.046
(0.029) (0.029)
16 years/older 0.076*** 0.076***
(0.025) (0.025)
Gender (male=1) 0.007 0.007
(0.018) (0.018)
Average household member age -0.009 -0.009 0.987 0.987
(0.006) (0.006) (0.014) (0.014)
Household size -0.013* -0.013* 0.021 0.021 1.062 1.062
(0.007) (0.007) (0.023) (0.023) (0.056) (0.055)
Household density 0.026*** 0.026*** 0.071*** 0.070*** 1.200*** 1.201***
(0.006) (0.006) (0.020) (0.020) (0.049) (0.048)
Highest education completed -0.004 -0.004 -0.002 -0.002 0.985 0.985
(0.004) (0.004) (0.012) (0.012) (0.028) (0.029)
Distance to health center -0.000 -0.000 0.004 0.004 0.999 0.999
(0.002) (0.002) (0.005) (0.005) (0.017) (0.016)
Livestock holding 0.008 0.009 0.032 0.030 1.085 1.086*
(0.007) (0.007) (0.023) (0.022) (0.055) (0.055)
Log of assets value 0.017 0.017 0.032 0.033 1.171 1.170
(0.013) (0.013) (0.041) (0.042) (0.119) (0.123)
Proportion of irrigated land 0.005 0.004 -0.059 -0.055 0.997 0.995
(0.029) (0.030) (0.090) (0.093) (0.232) (0.241)
Constant 0.057*** 0.057***
(0.053) (0.053)
Observation 1455 1455 302 302 302 302
Pseudo R-squared 0.04 0.04 0.11 0.11
Model Chi2 52.10 51.94 53.25 53.63 76.18 70.07
Model p-value 0.000 0.000 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1, Probit in average marginal effects and Poisson in incidence-rate ratios.
153
Table A7: Health effects of irrigation water use—Morbidity
(1) (2) (3) (4) (5) (6)
VARIABLES PROBIT PROBIT PROBIT PROBIT POISSON POISSON
Irrigation water use (dummy) 0.084*** 0.086*** 0.132*** 0.131*** 1.263*** 1.270***
(0.027) (0.027) (0.042) (0.042) (0.114) (0.114)
Water quality (1= no E.coli) -0.033 0.005 0.922
(0.030) (0.039) (0.085)
Water collection time (1=30min/less) -0.022 -0.026 -0.151*** -0.149*** 0.968 0.959
(0.036) (0.034) (0.053) (0.053) (0.104) (0.098)
Pit latrine (dummy) 0.074*** 0.074*** 0.064 0.064 1.247** 1.249**
(0.028) (0.028) (0.047) (0.047) (0.109) (0.110)
Age in years: reference (5—6)
7—9 years -0.116** -0.116**
(0.057) (0.058)
10—15 years -0.146*** -0.145***
(0.052) (0.052)
16 years/older -0.069 -0.068
(0.048) (0.048)
Gender (male=1) -0.049 -0.048
(0.031) (0.031)
Average household member age -0.009** -0.009** 0.963** 0.963**
(0.004) (0.004) (0.016) (0.016)
Household size 0.001 0.001 0.047*** 0.047*** 1.182*** 1.182***
(0.011) (0.011) (0.017) (0.018) (0.042) (0.041)
Household density 0.035*** 0.036*** 0.054** 0.053** 1.082** 1.086***
(0.012) (0.012) (0.022) (0.022) (0.035) (0.034)
Highest education completed -0.017*** -0.017*** -0.020*** -0.020*** 0.960** 0.959**
(0.005) (0.005) (0.007) (0.007) (0.017) (0.016)
Distance to health center 0.007** 0.008** 0.017*** 0.017*** 1.021** 1.022**
(0.004) (0.004) (0.006) (0.006) (0.011) (0.011)
Livestock holding 0.013 0.016 0.015 0.014 1.058* 1.067*
(0.011) (0.011) (0.017) (0.016) (0.035) (0.036)
Log of assets value 0.012 0.009 0.015 0.016 1.045 1.040
(0.018) (0.018) (0.032) (0.032) (0.057) (0.057)
Proportion of irrigated land -0.087** -0.095** -0.048 -0.046 0.714*** 0.700***
(0.043) (0.045) (0.056) (0.056) (0.091) (0.094)
Constant 0.548 0.544
(0.343) (0.338)
Observation 1455 1455 302 302 302 302
Pseudo R-squared 0.05 0.04 0.20 0.20
Model Chi2 108.21 100.65 71.91 71.80 238.45 211.23
Model p-value 0.000 0.000 0.000 0.000 0.000 0.000
Robust standard errors adjusted for clustering at village level in parentheses; Significance *** p<0.01, ** p<0.05, * p<0.1, Probit in average marginal effects and Poisson in incidence-rate ratios.
154
Table A8: Health effects of irrigation and irrigation water use—any illness except injury
(1) (2) VARIABLES POISSON SE POISSON SE
Irrigating households (dummy) 1.529*** 0.175
Irrigation water use (dummy) 1.299*** 0.123
Water quality (1=no E.coli) 0.872* 0.065 0.925 0.087
Water collection time (1=30min/less) 1.059 0.089 1.058 0.113
Pit latrine (dummy) 1.175** 0.087 1.262** 0.120
Average household member age 0.962*** 0.010 0.951*** 0.012
Household size 1.208*** 0.032 1.172*** 0.039
Household density 1.034 0.030 1.076** 0.032
Highest education completed 0.965*** 0.013 0.964** 0.017
Distance to health center 1.031*** 0.009 1.029** 0.011
Livestock holding 1.044 0.031 1.070** 0.036
Log of assets value 1.008 0.048 1.016 0.056
Proportion of irrigated land 0.672*** 0.094 0.710** 0.100
Constant 0.523 0.235 0.751 0.427
Observations 454 302
Model Chi2 326.04 223.55
Model p-value 0.000 0.000
Robust standard errors adjusted for clustering at village level; Significance *** p<0.01, ** p<0.05, * p<0.1, Regression confidents are in incidence-rate ratios.
155
Table A9: The association between irrigation and medical expenditure
(1) (2) VARIABLES Medical Expenditure
(last 2 months) Medical Expenditure
(last one year)
Irrigating households (dummy) 48.372*** 91.534*** (14.693) (30.300) Household size 12.294** 34.337*** (4.882) (7.428) Highest education completed 4.010 1.716 (2.816) (5.270) Log of non-farm income 2.385 10.822** (2.004) (4.530) Distance to health center 1.553 10.757** (2.278) (5.108) Constant -49.081 -83.992* (31.884) (49.484)
Observations 454 454 R-squared 0.05 0.07 Model F-stat 4.780 10.47 Model p-value 0.001 0.000
Robust standard errors adjusted for clustering at village level; Significance *** p<0.01, ** p<0.05, * p<0.1,
156
Table A10: Agricultural production estimates
(1) (2) VARIABLES Farm productivity SE Farm productivity SE
Irrigating households (dummy) 0.528*** 0.078 0.371*** 0.100
Agricultural equipment 0.045* 0.025 0.001 0.029
Total family male labor days 0.245*** 0.031 0.220*** 0.031
Total family female labor days 0.102** 0.045 0.077 0.049
Total hired male labor days 0.091*** 0.022 0.086*** 0.022
Total family female labor days 0.044** 0.018 0.034** 0.017
Total variable inputs 0.101*** 0.022 0.110*** 0.023
Proportion of irrigated land -0.284** 0.118 -0.196 0.122
Land quality index -0.191 0.118 -0.172 0.118
Livestock holding 0.080*** 0.021 0.061*** 0.021 Average distance to plot (minutes) -0.003* 0.002 -0.003** 0.002 Highest education completed -0.018* 0.009
Off farm labor participation (dummy) -0.074 0.069
Household size 0.043** 0.019 Assets value 0.087** 0.037 Altitude -1.518** 0.594 Constant 6.541*** 0.323 17.716*** 4.490
Observations 454 454 R-squared 0.54 0.56 Model F-stat 50.14 39.80 Model p-value 0.000 0.000
Robust standard errors; Significance *** p<0.01, ** p<0.05, * p<0.1, Variables are transformed into logarithm except level of education, household size and the dummy variables.
157
Appendix B
Participation Information Sheet and Informed Consent Form
Name of Project: Water, Sanitation, and Hygiene Nexus under a Multi-Use Water System: Synergies, Thresholds, and Trade-offs for a Better Health and Nutritional Outcomes in Ethiopia.
Part I: Information Sheet Introduction
My name is________ and I am conducting a survey for the AG-WATSAN-Nexus project. The center for Development Research (ZEF), University of Bonn, Germany in collaboration with the Ethiopian Economics Association (EEA), Addis Ababa, are now implementing the project to address basic health needs including household drinking water supply, sanitation, and knowledge and practices concerning hygiene in the selected villages of Fogera and Mecha Woredas. The project is funded by the Bill and Melinda Gates Foundation and this study will examine the synergies and trade-offs between domestic water supply (quality and quantity), sanitation, hygiene (WASH) and their linkage with irrigation and drainage systems.
As part of the planning process, we are conducting a survey of households in villages that are being considered for participation in the project. We have sampled certain households randomly to collect information about the household and the people living there; the current situation of water supply and sanitation facilities; knowledge and practices concerning hygiene; other health care and household practices; agricultural production, household consumption and expenditures; household time and labor use, and anthropometric measures for under-five children. You are one of them and invited to participate in this research and we would appreciate your contribution to our understanding on WASH in your area.
If you have any questions regarding this research project or anything you do not understand during the interview, you are free to ask and the interviewer will take time to explain to you to the best of his/her knowledge.
Purpose of the Research
The main purpose of this research is to analyze the impacts of use of different water sources and various sanitation facilities on health and their linkage with irrigation and drainage systems so that the findings will help the government or other stakeholders to plan better water supply, sanitation and essential health-care services for this areas.
Type of Research Intervention
This survey will involve different member of your household and the entire interview will take approximately not more than 60 minutes. However, households which have under-five children will be revisited every fortnightly for about five or six times (three months) to collect information about the prevalence of diarrheal diseases for last two-week period.
Voluntary Participation
Your participation in this research is completely voluntary. If for any reason you do not wish to participate, you can chose not to, and you can object to answering any specific question or questions in the questionnaire. There are no disadvantages to deciding not to participate or not to answer certain questions. It is entirely your choice whether to participate or not. However, we would greatly appreciate your cooperation.
Benefits of Participation
There will be no direct benefit to you, but your participation is likely to help us find out more about the present burden of water, sanitation and hygiene related diseases and eventually identify better strategies of linking domestic water uses for ‘WATSAN’ and irrigation and drainage systems to improve health and nutrition status in
158
your area. Apart from acknowledging your contribution in sparing time for us in answering the questions, we will also give you some present to compensate the loss of your time.
Confidentiality
Any information we obtain from you during the research will be kept strictly confidential and your answers will never be shared with anyone other than our project team and partner and will be kept private. Names and addresses of participants will not be included in the analysis or report, nor will information about a person’s household be shared with anyone else. The results of the study will be published in academic journals, policy briefing issues and as discussion papers.
Right to Refuse or Withdraw
You do not have to take part in this research if you do not wish to do so, and choosing to participate will not affect you in any way. You may stop participating in the interview at any time you wish. I will give you an opportunity at the end of the interview/discussion to review your remarks, and you can ask to modify or remove portions of those, if you do not agree with my notes or if I did not understand you correctly. You can also ask me any more questions about any part of the research study if you wish to.
Part II: Certificate of Consent
Participant Statement:
I have read (had someone read to me) this consent form. I have discussed with the research staff the information in this consent form. I have been given the opportunity to ask questions. I understand that I may refuse to participate in this study and that if I refuse to participate, this will not result in negative personal repercussions. I also understand that if, for any reason, I wish to stop participating, I will be free to do so. I have understood the purpose of the study and I am willing to participate in the interview.
Do you agree to participate?
Yes THEN BEGIN THE INTERVIEW. No DISCUSS THIS RESULT WITH YOUR SUPERVISOR AND GO TO THE NEXT HOUSEHOLD.
Name and Signature of Participant: ______________________________________ Date: ____________________
Statement by the researcher/person taking consent
I have accurately read out the information sheet to the potential participant, and to the best of my ability made sure that the participant understands. I confirm that the participant was given an opportunity to ask questions about the study, and all the questions asked by the participant have been answered to the best of my ability. I confirm that the individual has not been forced into giving consent, and the consent has been given freely and voluntarily.
Signature of Researcher /Person taking the consent: ______________________ Date: _____________________
REMARK: PREPARE 2 COPIES, AND ONE COPY SHOULD BE GIVEN TO THE PARTICIPANT!
Whom to Contact
If you feel you have been treated unfairly, or you have questions or concerns you may contact:
Ethiopian Economics Association, Ethiopian Economic Policy Research Institute (EEA/EEPRI),
P.O.Box: 34282
Addis Ababa, Ethiopia
Tel: 251-11-645-3200/3091
